diff --git a/cuda-oxide/crates/cuda-device/src/generated/async_copy.rs b/cuda-oxide/crates/cuda-device/src/generated/async_copy.rs index 03e050e5c..6005b2eec 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/async_copy.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/async_copy.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::async_copy` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/atomic.rs b/cuda-oxide/crates/cuda-device/src/generated/atomic.rs index 2dae6c039..294731d29 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/atomic.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/atomic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::atomic` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/bf16.rs b/cuda-oxide/crates/cuda-device/src/generated/bf16.rs index f452534a7..f7e9a219b 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/bf16.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/bf16.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::bf16` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/bf16x2.rs b/cuda-oxide/crates/cuda-device/src/generated/bf16x2.rs index 13faf64f3..ec3d6daaf 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/bf16x2.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/bf16x2.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::bf16x2` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/clc.rs b/cuda-oxide/crates/cuda-device/src/generated/clc.rs index 70cf41984..600b958d7 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/clc.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/clc.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::clc` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/cluster_barrier.rs b/cuda-oxide/crates/cuda-device/src/generated/cluster_barrier.rs index e227ffaee..40445e21b 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/cluster_barrier.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/cluster_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::cluster` to keep public paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/cluster_memory.rs b/cuda-oxide/crates/cuda-device/src/generated/cluster_memory.rs index fbff9c915..94b5fa04d 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/cluster_memory.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/cluster_memory.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::cluster` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/cluster_sreg.rs b/cuda-oxide/crates/cuda-device/src/generated/cluster_sreg.rs index 54ac3d92f..400c6d0de 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/cluster_sreg.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/cluster_sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // This file is included inside `cuda_device::cluster`. diff --git a/cuda-oxide/crates/cuda-device/src/generated/convert.rs b/cuda-oxide/crates/cuda-device/src/generated/convert.rs index 883d367a2..14d972971 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/convert.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/convert.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::convert`. diff --git a/cuda-oxide/crates/cuda-device/src/generated/counted_barrier.rs b/cuda-oxide/crates/cuda-device/src/generated/counted_barrier.rs index a51ac2a76..26128f3f2 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/counted_barrier.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/counted_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::barrier` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/debug_control.rs b/cuda-oxide/crates/cuda-device/src/generated/debug_control.rs index fc4f94874..5ccf65104 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/debug_control.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/debug_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::debug` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/debug_sreg.rs b/cuda-oxide/crates/cuda-device/src/generated/debug_sreg.rs index 1c38bef61..6271828e9 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/debug_sreg.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/debug_sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::debug` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/dotprod.rs b/cuda-oxide/crates/cuda-device/src/generated/dotprod.rs index 11bcebced..d36bd2e52 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/dotprod.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/dotprod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::dotprod` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/f16.rs b/cuda-oxide/crates/cuda-device/src/generated/f16.rs index 37bb5d205..dd1f10a87 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/f16.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/f16.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::f16` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/f16x2.rs b/cuda-oxide/crates/cuda-device/src/generated/f16x2.rs index e0ba7d115..dbaa705cd 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/f16x2.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/f16x2.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::f16x2` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/f32x2.rs b/cuda-oxide/crates/cuda-device/src/generated/f32x2.rs index 82010ea7d..6ba54ef4b 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/f32x2.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/f32x2.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::f32x2` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/fence.rs b/cuda-oxide/crates/cuda-device/src/generated/fence.rs index 7025a6e77..1a932ef84 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/fence.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/fence.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // This file is included inside `cuda_device::fence` so existing paths stay stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/float.rs b/cuda-oxide/crates/cuda-device/src/generated/float.rs index bb17949ce..16a109f31 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/float.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/float.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::float`. diff --git a/cuda-oxide/crates/cuda-device/src/generated/grid_dependency.rs b/cuda-oxide/crates/cuda-device/src/generated/grid_dependency.rs index 7660a7419..32d74cbc0 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/grid_dependency.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/grid_dependency.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::grid` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/grid_sreg.rs b/cuda-oxide/crates/cuda-device/src/generated/grid_sreg.rs index e80f63b6c..aaf16a69b 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/grid_sreg.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/grid_sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::grid` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/i16x2.rs b/cuda-oxide/crates/cuda-device/src/generated/i16x2.rs index 8079ed880..e52f27107 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/i16x2.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/i16x2.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::i16x2` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/int.rs b/cuda-oxide/crates/cuda-device/src/generated/int.rs index 91e59c67f..2a130f9a7 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/int.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/int.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::int` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/ldmatrix.rs b/cuda-oxide/crates/cuda-device/src/generated/ldmatrix.rs index 15c90d0ae..7213f1a85 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/ldmatrix.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/ldmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::wmma` to keep public paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/mbarrier_basic.rs b/cuda-oxide/crates/cuda-device/src/generated/mbarrier_basic.rs index 42420ffe4..ac64208cd 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/mbarrier_basic.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/mbarrier_basic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::barrier` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/mbarrier_extended.rs b/cuda-oxide/crates/cuda-device/src/generated/mbarrier_extended.rs index 499be8e9e..ab37209b1 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/mbarrier_extended.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/mbarrier_extended.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::barrier` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/movmatrix.rs b/cuda-oxide/crates/cuda-device/src/generated/movmatrix.rs index 958ac2da4..6cafa5dff 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/movmatrix.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/movmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::wmma` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/prmt.rs b/cuda-oxide/crates/cuda-device/src/generated/prmt.rs index ebbac73f0..7068c9f98 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/prmt.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/prmt.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::prmt` to keep the public paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/register_control.rs b/cuda-oxide/crates/cuda-device/src/generated/register_control.rs index f45cc3ae7..dc19ccc2d 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/register_control.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/register_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::thread` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/register_mma.rs b/cuda-oxide/crates/cuda-device/src/generated/register_mma.rs index 654e3954e..8b99762cc 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/register_mma.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/register_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::wmma` to keep public paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/shared_sreg.rs b/cuda-oxide/crates/cuda-device/src/generated/shared_sreg.rs index 8e2dd4cde..7cc59247f 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/shared_sreg.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/shared_sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::shared` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/sparse_mma.rs b/cuda-oxide/crates/cuda-device/src/generated/sparse_mma.rs index 80526e716..5e69b8f2f 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/sparse_mma.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/sparse_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::wmma` to keep public paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/sreg.rs b/cuda-oxide/crates/cuda-device/src/generated/sreg.rs index 6f178f601..dfde583a2 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/sreg.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // This file is included lexically inside `cuda_device::thread` so existing diff --git a/cuda-oxide/crates/cuda-device/src/generated/stmatrix.rs b/cuda-oxide/crates/cuda-device/src/generated/stmatrix.rs index 25577a893..433011ffc 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/stmatrix.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/stmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::tcgen05` to keep public paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/tcgen05.rs b/cuda-oxide/crates/cuda-device/src/generated/tcgen05.rs index 508f5eada..e1a13e41b 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/tcgen05.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/tcgen05.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::tcgen05` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/tma.rs b/cuda-oxide/crates/cuda-device/src/generated/tma.rs index 8800c3a64..93cb9d720 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/tma.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/tma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::tma` to keep its public API stable. @@ -233,6 +233,23 @@ pub unsafe fn cp_async_bulk_tensor_1d_g2s( unreachable!("cp_async_bulk_tensor_1d_g2s called outside CUDA kernel context") } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// # Safety +/// `dst`, `tensor_map`, and `barrier` must remain valid until the copy completes. +/// `barrier` must point into the issuing CTA's shared memory, as for the other +/// `mbarrier` operations. `dst` is a [`SharedPtr`], which points there by its type. +#[inline(never)] +pub unsafe fn cp_async_bulk_tensor_1d_g2s_cta( + dst: SharedPtr, + tensor_map: *const TmaDescriptor, + coord0: i32, + barrier: *mut Barrier, +) { + let _ = (dst, tensor_map, coord0, barrier); + unreachable!("cp_async_bulk_tensor_1d_g2s_cta called outside CUDA kernel context") +} + /// Starts a TMA tile copy from shared to global memory. /// /// # Safety @@ -263,6 +280,24 @@ pub unsafe fn cp_async_bulk_tensor_2d_g2s( unreachable!("cp_async_bulk_tensor_2d_g2s called outside CUDA kernel context") } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// # Safety +/// `dst`, `tensor_map`, and `barrier` must remain valid until the copy completes. +/// `barrier` must point into the issuing CTA's shared memory, as for the other +/// `mbarrier` operations. `dst` is a [`SharedPtr`], which points there by its type. +#[inline(never)] +pub unsafe fn cp_async_bulk_tensor_2d_g2s_cta( + dst: SharedPtr, + tensor_map: *const TmaDescriptor, + coord0: i32, + coord1: i32, + barrier: *mut Barrier, +) { + let _ = (dst, tensor_map, coord0, coord1, barrier); + unreachable!("cp_async_bulk_tensor_2d_g2s_cta called outside CUDA kernel context") +} + /// Starts a multicast TMA tile copy from global to cluster shared memory. /// /// # Safety @@ -329,6 +364,25 @@ pub unsafe fn cp_async_bulk_tensor_3d_g2s( unreachable!("cp_async_bulk_tensor_3d_g2s called outside CUDA kernel context") } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// # Safety +/// `dst`, `tensor_map`, and `barrier` must remain valid until the copy completes. +/// `barrier` must point into the issuing CTA's shared memory, as for the other +/// `mbarrier` operations. `dst` is a [`SharedPtr`], which points there by its type. +#[inline(never)] +pub unsafe fn cp_async_bulk_tensor_3d_g2s_cta( + dst: SharedPtr, + tensor_map: *const TmaDescriptor, + coord0: i32, + coord1: i32, + coord2: i32, + barrier: *mut Barrier, +) { + let _ = (dst, tensor_map, coord0, coord1, coord2, barrier); + unreachable!("cp_async_bulk_tensor_3d_g2s_cta called outside CUDA kernel context") +} + /// Starts a TMA tile copy from shared to global memory. /// /// # Safety @@ -363,6 +417,26 @@ pub unsafe fn cp_async_bulk_tensor_4d_g2s( unreachable!("cp_async_bulk_tensor_4d_g2s called outside CUDA kernel context") } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// # Safety +/// `dst`, `tensor_map`, and `barrier` must remain valid until the copy completes. +/// `barrier` must point into the issuing CTA's shared memory, as for the other +/// `mbarrier` operations. `dst` is a [`SharedPtr`], which points there by its type. +#[inline(never)] +pub unsafe fn cp_async_bulk_tensor_4d_g2s_cta( + dst: SharedPtr, + tensor_map: *const TmaDescriptor, + coord0: i32, + coord1: i32, + coord2: i32, + coord3: i32, + barrier: *mut Barrier, +) { + let _ = (dst, tensor_map, coord0, coord1, coord2, coord3, barrier); + unreachable!("cp_async_bulk_tensor_4d_g2s_cta called outside CUDA kernel context") +} + /// Starts a TMA tile copy from shared to global memory. /// /// # Safety @@ -402,6 +476,30 @@ pub unsafe fn cp_async_bulk_tensor_5d_g2s( unreachable!("cp_async_bulk_tensor_5d_g2s called outside CUDA kernel context") } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// # Safety +/// `dst`, `tensor_map`, and `barrier` must remain valid until the copy completes. +/// `barrier` must point into the issuing CTA's shared memory, as for the other +/// `mbarrier` operations. `dst` is a [`SharedPtr`], which points there by its type. +#[inline(never)] +#[allow(clippy::too_many_arguments)] +pub unsafe fn cp_async_bulk_tensor_5d_g2s_cta( + dst: SharedPtr, + tensor_map: *const TmaDescriptor, + coord0: i32, + coord1: i32, + coord2: i32, + coord3: i32, + coord4: i32, + barrier: *mut Barrier, +) { + let _ = ( + dst, tensor_map, coord0, coord1, coord2, coord3, coord4, barrier, + ); + unreachable!("cp_async_bulk_tensor_5d_g2s_cta called outside CUDA kernel context") +} + /// Starts a TMA tile copy from shared to global memory. /// /// # Safety diff --git a/cuda-oxide/crates/cuda-device/src/generated/warp_sreg.rs b/cuda-oxide/crates/cuda-device/src/generated/warp_sreg.rs index 8f1aa708c..557554a1a 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/warp_sreg.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/warp_sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::warp` to keep existing paths stable. diff --git a/cuda-oxide/crates/cuda-device/src/generated/wgmma_control.rs b/cuda-oxide/crates/cuda-device/src/generated/wgmma_control.rs index 750c0d2d4..ba8dd5c16 100644 --- a/cuda-oxide/crates/cuda-device/src/generated/wgmma_control.rs +++ b/cuda-oxide/crates/cuda-device/src/generated/wgmma_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. // Included inside `cuda_device::wgmma` to keep its public API stable. diff --git a/cuda-oxide/crates/cuda-device/src/lib.rs b/cuda-oxide/crates/cuda-device/src/lib.rs index a0b85677c..67ee46790 100644 --- a/cuda-oxide/crates/cuda-device/src/lib.rs +++ b/cuda-oxide/crates/cuda-device/src/lib.rs @@ -82,7 +82,7 @@ pub use cusimd::{CuSimd, Float2, Float4, TmemRegs4, TmemRegs32}; pub use disjoint::{__LaunchContractDisjointSlice, __LaunchContractDisjointSliceAbi}; pub use disjoint::{DisjointSlice, SpaceLayout}; pub use fence::*; -pub use shared::{DynamicSharedArray, SharedArray}; +pub use shared::{DynamicSharedArray, SharedArray, SharedPtr}; pub use tcgen05::{ TensorMemoryHandle, TmemAddress, TmemDeallocated, TmemF32x4, TmemF32x32, TmemGuard, TmemReady, TmemUninit, diff --git a/cuda-oxide/crates/cuda-device/src/shared.rs b/cuda-oxide/crates/cuda-device/src/shared.rs index a5e459858..790e24078 100644 --- a/cuda-oxide/crates/cuda-device/src/shared.rs +++ b/cuda-oxide/crates/cuda-device/src/shared.rs @@ -233,6 +233,34 @@ impl SharedArray { let _ = shared; unreachable!("SharedArray::as_raw_mut_ptr called outside CUDA kernel context") } + + /// Returns a [`SharedPtr`] to the first element of a shared array. + /// + /// Unlike [`Self::as_raw_mut_ptr`], which returns a generic pointer, the + /// result keeps the array's CTA-shared address space in its type, so it + /// stays a shared-memory pointer when it is passed to helper functions. + /// + /// # Safety + /// + /// `shared` must point to a `static mut SharedArray` in the current CUDA + /// kernel, normally obtained with `&raw mut`. The returned pointer is valid + /// only within that kernel. This requirement is what makes the type's + /// claim true: a `SharedPtr` built from any other pointer hands + /// [`SharedPtr::addr_u32`] an address outside the shared window, where the + /// conversion yields an unspecified value. Callers must keep concurrent + /// accesses disjoint and use the CUDA synchronization required before + /// reading data written by another thread. + /// + /// ```rust,ignore + /// static mut TILE: SharedArray = SharedArray::UNINIT; + /// let tile: SharedPtr = unsafe { SharedArray::shared_ptr(&raw mut TILE) }; + /// ``` + #[inline(always)] + pub unsafe fn shared_ptr(shared: *mut Self) -> SharedPtr { + SharedPtr { + ptr: shared as *mut SharedSlot, + } + } } impl Index for SharedArray { @@ -520,6 +548,24 @@ impl DynamicSharedArray { let _ = byte_offset; unreachable!("DynamicSharedArray::offset called outside CUDA kernel context") } + + /// Like [`Self::get`], but returns a [`SharedPtr`], which keeps the + /// CTA-shared address space in its type. + #[inline(always)] + pub fn shared_ptr() -> SharedPtr { + SharedPtr { + ptr: Self::get() as *mut SharedSlot, + } + } + + /// Like [`Self::offset`], but returns a [`SharedPtr`], which keeps the + /// CTA-shared address space in its type. + #[inline(always)] + pub fn shared_ptr_at(byte_offset: usize) -> SharedPtr { + SharedPtr { + ptr: Self::offset(byte_offset) as *mut SharedSlot, + } + } } /// Convert a generic-address pointer into its raw `.shared` window offset. @@ -573,4 +619,136 @@ pub unsafe fn cvta_generic_to_shared_u32(ptr: *const u8) -> u32 { unreachable!("cvta_generic_to_shared_u32 called outside CUDA kernel context") } +// ============================================================================ +// SharedPtr - a pointer into the issuing CTA's shared memory, typed as such +// ============================================================================ + +/// The pointee of [`SharedPtr`]'s field. A pointer to `SharedSlot` is a +/// pointer into CTA-shared memory (`addrspace(3)`) by its type, as a pointer +/// to [`SharedArray`] is; `repr(transparent)` gives it `T`'s size, so pointer +/// arithmetic strides by `T`. +#[doc(hidden)] +#[repr(transparent)] +pub struct SharedSlot(T); + +/// A raw pointer into the shared memory of the issuing CTA (`.shared::cta`). +/// +/// A `*mut T` in device code is a generic pointer, which may address global, +/// local or shared memory, including another CTA's shared memory in the +/// cluster. `SharedPtr` instead carries its address space in its type: the +/// compiler knows it is a CTA-shared pointer wherever it goes, including +/// through calls to helper functions, and emits the `.shared::cta` forms of +/// instructions that take it without converting a generic address. +/// +/// It is created from shared memory the CTA owns, with the `unsafe` +/// [`SharedArray::shared_ptr`] or the safe [`DynamicSharedArray::shared_ptr`], +/// or from a generic pointer with the `unsafe` [`SharedPtr::from_raw`]. Like a raw +/// pointer, it may dangle; dereferencing it goes through [`SharedPtr::as_ptr`] +/// and `unsafe` code. +/// +/// Pointers from [`crate::cluster::map_shared_rank`], which may name another +/// CTA's shared memory, are not `SharedPtr`s. +/// +/// # Example +/// +/// ```rust,ignore +/// static mut TILE: SharedArray = SharedArray::UNINIT; +/// +/// fn store(tile: SharedPtr, i: usize, value: f32) { +/// // `tile` is still a shared pointer here, although `store` is a +/// // separate function. +/// unsafe { tile.add(i).as_ptr().write(value) }; +/// } +/// +/// let tile = unsafe { SharedArray::shared_ptr(&raw mut TILE) }; +/// store(tile, thread::threadIdx_x() as usize, 1.0); +/// ``` +#[repr(transparent)] +pub struct SharedPtr { + ptr: *mut SharedSlot, +} + +impl Clone for SharedPtr { + #[inline(always)] + fn clone(&self) -> Self { + *self + } +} + +impl Copy for SharedPtr {} + +impl SharedPtr { + /// Converts a generic pointer to a `SharedPtr`, as CUDA C++'s + /// `__cvta_generic_to_shared` or CuTe's `make_smem_ptr` do. + /// + /// # Safety + /// + /// `ptr` must point into the shared memory of the issuing CTA. A pointer + /// to global or local memory, or to another CTA's shared memory, gives an + /// unspecified address. + #[inline(always)] + pub unsafe fn from_raw(ptr: *mut T) -> Self { + Self { + ptr: ptr as *mut SharedSlot, + } + } + + /// The generic pointer to the same memory, for ordinary loads and stores + /// and for APIs that take raw pointers. + #[inline(always)] + pub fn as_ptr(self) -> *mut T { + self.ptr as *mut T + } + + /// The pointer's 32-bit `.shared::cta` address, as taken by the + /// `_shared_u32` forms of CTA-shared instructions. + #[inline(always)] + pub fn addr_u32(self) -> u32 { + // SAFETY: `self` points into the issuing CTA's shared memory. + unsafe { cvta_generic_to_shared_u32(self.as_ptr() as *const u8) } + } + + /// Reinterprets the pointee type, keeping the address space. + #[inline(always)] + pub fn cast(self) -> SharedPtr { + SharedPtr { + ptr: self.ptr as *mut SharedSlot, + } + } + + /// The pointer `count` elements of `T` further on, as `<*mut T>::add`. + /// + /// # Safety + /// + /// As for `<*mut T>::add`. + #[inline(always)] + pub unsafe fn add(self, count: usize) -> Self { + Self { + ptr: unsafe { self.ptr.add(count) }, + } + } + + /// The pointer `count` elements of `T` away, as `<*mut T>::offset`. + /// + /// # Safety + /// + /// As for `<*mut T>::offset`. + #[inline(always)] + pub unsafe fn offset(self, count: isize) -> Self { + Self { + ptr: unsafe { self.ptr.offset(count) }, + } + } + + /// The pointer `count` bytes further on, as `<*mut T>::byte_add`. + /// + /// # Safety + /// + /// As for `<*mut T>::byte_add`. + #[inline(always)] + pub unsafe fn byte_add(self, count: usize) -> Self { + unsafe { self.cast::().add(count).cast::() } + } +} + include!("generated/shared_sreg.rs"); diff --git a/cuda-oxide/crates/cuda-device/src/tma.rs b/cuda-oxide/crates/cuda-device/src/tma.rs index 3532cab7e..fdbc2fb0b 100644 --- a/cuda-oxide/crates/cuda-device/src/tma.rs +++ b/cuda-oxide/crates/cuda-device/src/tma.rs @@ -113,6 +113,7 @@ //! - **sm_120 (Blackwell)**: Latest TMA capabilities use crate::barrier::Barrier; +use crate::shared::SharedPtr; // ============================================================================= // TMA Descriptor Type diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/model/contracts/tma.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/model/contracts/tma.rs index d4690333e..5270e609c 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/model/contracts/tma.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/model/contracts/tma.rs @@ -30,12 +30,17 @@ impl Tma { #[serde(rename_all = "snake_case")] pub enum TmaOperation { G2sTile1d, + G2sCtaTile1d, G2sTile2d, + G2sCtaTile2d, G2sTile2dMulticast, G2sTile2dMulticastCg2, G2sTile3d, + G2sCtaTile3d, G2sTile4d, + G2sCtaTile4d, G2sTile5d, + G2sCtaTile5d, S2gTile1d, S2gTile2d, S2gTile3d, @@ -82,16 +87,28 @@ pub enum TmaOperation { } impl TmaOperation { + pub const fn is_cta_g2s(self) -> bool { + matches!( + self, + Self::G2sCtaTile1d + | Self::G2sCtaTile2d + | Self::G2sCtaTile3d + | Self::G2sCtaTile4d + | Self::G2sCtaTile5d + ) + } + pub const fn dimensions(self) -> Option { match self { - Self::G2sTile1d | Self::S2gTile1d => Some(1), + Self::G2sTile1d | Self::G2sCtaTile1d | Self::S2gTile1d => Some(1), Self::G2sTile2d + | Self::G2sCtaTile2d | Self::G2sTile2dMulticast | Self::G2sTile2dMulticastCg2 | Self::S2gTile2d => Some(2), - Self::G2sTile3d | Self::S2gTile3d => Some(3), - Self::G2sTile4d | Self::S2gTile4d => Some(4), - Self::G2sTile5d | Self::S2gTile5d => Some(5), + Self::G2sTile3d | Self::G2sCtaTile3d | Self::S2gTile3d => Some(3), + Self::G2sTile4d | Self::G2sCtaTile4d | Self::S2gTile4d => Some(4), + Self::G2sTile5d | Self::G2sCtaTile5d | Self::S2gTile5d => Some(5), Self::Reduce | Self::CommitGroup | Self::WaitGroup diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/model/overlay.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/model/overlay.rs index bea0d9e44..2a7dd95c5 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/model/overlay.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/model/overlay.rs @@ -324,6 +324,12 @@ pub struct ClcAdmissionVariant { pub struct TmaAdmission { pub llvm_evidence_profile: String, pub libnvvm_evidence_profile: String, + /// Evidence profiles for the CTA-local G2S variants. A shard that lists + /// those variants must name both; an older shard without them parses. + #[serde(default)] + pub cta_llvm_evidence_profile: Option, + #[serde(default)] + pub cta_libnvvm_evidence_profile: Option, #[serde(default)] pub reduce_llvm_evidence_profile: Option, #[serde(default)] @@ -343,6 +349,16 @@ pub struct TmaAdmissionVariant { pub operation: TmaOperation, } +impl TmaAdmission { + /// Whether the shard lists the CTA-local G2S variants. Those variants + /// carry their own evidence profiles and need overlay shard schema 65. + pub fn admits_cta_g2s(&self) -> bool { + self.variants + .iter() + .any(|variant| variant.operation.is_cta_g2s()) + } +} + /// One reviewed tensor-reduction operation and its reserved ABI ID. #[derive(Debug, Clone, Deserialize)] #[serde(deny_unknown_fields)] diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/compat.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/compat.rs index 727569842..a396dc204 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/compat.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/compat.rs @@ -913,12 +913,17 @@ pub(super) fn render_compat_tma(catalog: &CatalogFile, hash: &str) -> String { let is_g2s = matches!( operation, TmaOperation::G2sTile1d + | TmaOperation::G2sCtaTile1d | TmaOperation::G2sTile2d + | TmaOperation::G2sCtaTile2d | TmaOperation::G2sTile2dMulticast | TmaOperation::G2sTile2dMulticastCg2 | TmaOperation::G2sTile3d + | TmaOperation::G2sCtaTile3d | TmaOperation::G2sTile4d + | TmaOperation::G2sCtaTile4d | TmaOperation::G2sTile5d + | TmaOperation::G2sCtaTile5d ); let is_s2g = matches!( operation, @@ -935,6 +940,11 @@ pub(super) fn render_compat_tma(catalog: &CatalogFile, hash: &str) -> String { output.push_str( "/// `dst`, `tensor_map`, and `barrier` must remain valid until the copy completes.\n", ); + if operation.is_cta_g2s() { + output.push_str( + "/// `barrier` must point into the issuing CTA's shared memory, as for the other\n/// `mbarrier` operations. `dst` is a [`SharedPtr`], which points there by its type.\n", + ); + } } else if is_s2g { output.push_str( "/// `src` and `tensor_map` must remain valid until the committed copy group completes.\n", @@ -961,8 +971,15 @@ pub(super) fn render_compat_tma(catalog: &CatalogFile, hash: &str) -> String { output.push_str("#[inline(never)]\n"); if is_g2s { let dimensions = dimensions.unwrap(); + // The CTA-local forms take a typed CTA-shared destination; the + // raw ABI keeps `*mut u8`, as `*mut Barrier` keeps `*mut u64`. + let destination = if operation.is_cta_g2s() { + "dst: SharedPtr" + } else { + "dst: *mut u8" + }; let mut arguments = vec![ - "dst: *mut u8".to_owned(), + destination.to_owned(), "tensor_map: *const TmaDescriptor".to_owned(), ]; let mut values = vec!["dst".to_owned(), "tensor_map".to_owned()]; diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/importer.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/importer.rs index 68f95d4e2..26883fe97 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/importer.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/importer.rs @@ -2402,6 +2402,18 @@ fn tma_arms(catalog: &CatalogFile) -> String { ) .unwrap(); } + TmaOperation::G2sCtaTile1d + | TmaOperation::G2sCtaTile2d + | TmaOperation::G2sCtaTile3d + | TmaOperation::G2sCtaTile4d + | TmaOperation::G2sCtaTile5d => { + writeln!( + output, + " Ok(Some(super::super::tma::emit_tma_g2s_cta(\n ctx, body, args, target, block_ptr, prev_op, value_map, block_map, loc, {}, {marker:?},\n )?))", + operation.dimensions().unwrap() + ) + .unwrap(); + } TmaOperation::G2sTile2dMulticast => { writeln!( output, diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/lowering.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/lowering.rs index 00efa16a5..09c0b9f03 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/lowering.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/lowering.rs @@ -1520,6 +1520,18 @@ fn tma_impls(catalog: &CatalogFile) -> String { ) .unwrap(); } + TmaOperation::G2sCtaTile1d + | TmaOperation::G2sCtaTile2d + | TmaOperation::G2sCtaTile3d + | TmaOperation::G2sCtaTile4d + | TmaOperation::G2sCtaTile5d => { + writeln!( + output, + " convert_g2s_cta(ctx, rewriter, self.get_operation(), operands_info, {})", + operation.dimensions().unwrap() + ) + .unwrap(); + } TmaOperation::G2sTile2dMulticast => output.push_str( " convert_g2s(ctx, rewriter, self.get_operation(), operands_info, 2, true)\n", ), @@ -2100,6 +2112,7 @@ const LOWERING_INTRINSIC_HELPERS: &[(&str, &str)] = &[ ("ReduceConfig", "tma::ReduceConfig"), ("convert_control", "tma::convert_control"), ("convert_g2s", "tma::convert_g2s"), + ("convert_g2s_cta", "tma::convert_g2s_cta"), ( "convert_g2s_multicast_cg2", "tma::convert_g2s_multicast_cg2", diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/probes.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/probes.rs index 3c7858866..980c03489 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/probes.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/probes.rs @@ -84,12 +84,17 @@ fn render_tma_probe(catalog: &CatalogFile, record: &CatalogIntrinsic, hash: &str let is_g2s = matches!( operation, TmaOperation::G2sTile1d + | TmaOperation::G2sCtaTile1d | TmaOperation::G2sTile2d + | TmaOperation::G2sCtaTile2d | TmaOperation::G2sTile2dMulticast | TmaOperation::G2sTile2dMulticastCg2 | TmaOperation::G2sTile3d + | TmaOperation::G2sCtaTile3d | TmaOperation::G2sTile4d + | TmaOperation::G2sCtaTile4d | TmaOperation::G2sTile5d + | TmaOperation::G2sCtaTile5d ); let coordinates = std::iter::repeat_n("i32", dimensions) .collect::>() @@ -103,6 +108,14 @@ fn render_tma_probe(catalog: &CatalogFile, record: &CatalogIntrinsic, hash: &str .collect::>() .join(", "); if is_g2s { + let cta = matches!( + operation, + TmaOperation::G2sCtaTile1d + | TmaOperation::G2sCtaTile2d + | TmaOperation::G2sCtaTile3d + | TmaOperation::G2sCtaTile4d + | TmaOperation::G2sCtaTile5d + ); let multicast = matches!( operation, TmaOperation::G2sTile2dMulticast | TmaOperation::G2sTile2dMulticastCg2 @@ -112,11 +125,19 @@ fn render_tma_probe(catalog: &CatalogFile, record: &CatalogIntrinsic, hash: &str } else { format!(", {coordinates}") }; - writeln!( - output, - "declare void @{symbol}(ptr addrspace(7), ptr addrspace(3), ptr{declaration_coordinates}, i16, i64, i1, i1, i32) #0\n" - ) - .unwrap(); + if cta { + writeln!( + output, + "declare void @{symbol}(ptr addrspace(3), ptr addrspace(3), ptr{declaration_coordinates}, i64, i1) #0\n" + ) + .unwrap(); + } else { + writeln!( + output, + "declare void @{symbol}(ptr addrspace(7), ptr addrspace(3), ptr{declaration_coordinates}, i16, i64, i1, i1, i32) #0\n" + ) + .unwrap(); + } let parameters = if coordinate_parameters.is_empty() { String::new() } else { @@ -129,9 +150,15 @@ fn render_tma_probe(catalog: &CatalogFile, record: &CatalogIntrinsic, hash: &str record.id ) .unwrap(); - output.push_str( - " %dst = addrspacecast ptr %dst_generic to ptr addrspace(7)\n %barrier = addrspacecast ptr %barrier_generic to ptr addrspace(3)\n", - ); + if cta { + output.push_str( + " %dst = addrspacecast ptr %dst_generic to ptr addrspace(3)\n %barrier = addrspacecast ptr %barrier_generic to ptr addrspace(3)\n", + ); + } else { + output.push_str( + " %dst = addrspacecast ptr %dst_generic to ptr addrspace(7)\n %barrier = addrspacecast ptr %barrier_generic to ptr addrspace(3)\n", + ); + } let arguments = if coordinate_arguments.is_empty() { String::new() } else { @@ -143,11 +170,19 @@ fn render_tma_probe(catalog: &CatalogFile, record: &CatalogIntrinsic, hash: &str } else { 0 }; - writeln!( - output, - " call void @{symbol}(ptr addrspace(7) %dst, ptr addrspace(3) %barrier, ptr %tensor_map{arguments}, i16 {mask}, i64 0, i1 {multicast}, i1 false, i32 {group}) #0" - ) - .unwrap(); + if cta { + writeln!( + output, + " call void @{symbol}(ptr addrspace(3) %dst, ptr addrspace(3) %barrier, ptr %tensor_map{arguments}, i64 0, i1 false) #0" + ) + .unwrap(); + } else { + writeln!( + output, + " call void @{symbol}(ptr addrspace(7) %dst, ptr addrspace(3) %barrier, ptr %tensor_map{arguments}, i16 {mask}, i64 0, i1 {multicast}, i1 false, i32 {group}) #0" + ) + .unwrap(); + } } else { let declaration_coordinates = if coordinates.is_empty() { String::new() diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/raw_abi.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/raw_abi.rs index 6477e238c..20fac09b2 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/raw_abi.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/raw_abi.rs @@ -330,6 +330,15 @@ pub(super) fn render_raw_abi(catalog: &CatalogFile, hash: &str) -> Result + { + output.push_str( + "/// The destination and barrier must lie in the issuing CTA's own shared memory, and the tensor map must be a live descriptor for this dimensionality.\n\ + /// A generic destination or barrier pointer is converted with `cvta.to.shared`, so a cluster-shared pointer from `map_shared_rank` would yield a wrong local offset; copies into another CTA's shared memory use the cluster form.\n\ + /// Keep every object alive until the asynchronous copy completes. Only the designated issuing thread may start this transfer.\n", + ) + } TmaAdapter::G2sPointersCoordinatesBarrierInjectDefaults | TmaAdapter::G2sPointersCoordinatesBarrierMaskInjectDefaults => output .push_str( diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/mma.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/mma.rs index 561f5e421..b45c3afba 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/mma.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/mma.rs @@ -345,7 +345,7 @@ fn register_mma_rendering_preserves_apis_order_convergence_and_variants() { let repo_root = Path::new(env!("CARGO_MANIFEST_DIR")).join("../.."); let catalog = crate::resolve::resolve(&repo_root).unwrap(); validate_renderable(&catalog).unwrap(); - assert_eq!(catalog.intrinsics.len(), 1033); + assert_eq!(catalog.intrinsics.len(), 1038); let records: Vec<_> = register_mmas(&catalog).collect(); assert_eq!(records.len(), 154); let generated_records = records diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/mod.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/mod.rs index 729db95fa..9afc2769f 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/mod.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/mod.rs @@ -42,7 +42,7 @@ fn catalog_with_clc() -> CatalogFile { fn catalog_with_tma() -> CatalogFile { let repo_root = Path::new(env!("CARGO_MANIFEST_DIR")).join("../.."); let catalog = crate::resolve::test_catalog_with_tma(&repo_root).unwrap(); - assert_eq!(tma_intrinsics(&catalog).count(), 111); + assert_eq!(tma_intrinsics(&catalog).count(), 116); catalog } diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/tensor.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/tensor.rs index d431282c5..dec130798 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/tensor.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/tests/tensor.rs @@ -419,7 +419,13 @@ fn tma_rendering_preserves_api_and_injects_backend_defaults() { assert!(compatibility.contains( "pub unsafe fn cp_async_bulk_tensor_reduce_xor_im2col_5d(src: *const u8, tensor_map: *const TmaDescriptor, coord0: i32, coord1: i32, coord2: i32, coord3: i32, coord4: i32)" )); + assert!(compatibility.contains( + "pub unsafe fn cp_async_bulk_tensor_2d_g2s_cta(dst: SharedPtr, tensor_map: *const TmaDescriptor, coord0: i32, coord1: i32, barrier: *mut Barrier)" + )); for dimensions in 1..=5 { + assert!(compatibility.contains(&format!( + "pub unsafe fn cp_async_bulk_tensor_{dimensions}d_g2s_cta(dst: SharedPtr," + ))); assert!(compatibility.contains(&format!( "pub unsafe fn cp_async_bulk_prefetch_tensor_{dimensions}d_l2(" ))); @@ -437,8 +443,8 @@ fn tma_rendering_preserves_api_and_injects_backend_defaults() { assert!(compatibility.contains("pub fn fence_proxy_tensormap_generic_release_system()")); let dialect = render_dialect_tma(&catalog, "test-hash"); - assert_eq!(dialect.matches("pub struct ").count(), 111); - assert_eq!(dialect.matches("NResultsInterface<0>").count(), 111); + assert_eq!(dialect.matches("pub struct ").count(), 116); + assert_eq!(dialect.matches("NResultsInterface<0>").count(), 116); assert!(dialect.contains("NOpdsInterface<10>")); assert!(dialect.contains("CpAsyncBulkWaitGroupReadOp::register(ctx)")); assert!(dialect.contains("CpAsyncBulkPrefetchTensorGather4TwoDimensionalL2Op::register(ctx)")); diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/validate.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/validate.rs index f78c55a7b..defa726c9 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/render/validate.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/render/validate.rs @@ -733,10 +733,15 @@ pub(super) fn validate_renderable(catalog: &CatalogFile) -> Result<()> { && match (tma.operation, tma.adapter) { ( TmaOperation::G2sTile1d + | TmaOperation::G2sCtaTile1d | TmaOperation::G2sTile2d + | TmaOperation::G2sCtaTile2d | TmaOperation::G2sTile3d + | TmaOperation::G2sCtaTile3d | TmaOperation::G2sTile4d - | TmaOperation::G2sTile5d, + | TmaOperation::G2sCtaTile4d + | TmaOperation::G2sTile5d + | TmaOperation::G2sCtaTile5d, TmaAdapter::G2sPointersCoordinatesBarrierInjectDefaults, ) => { !record.rust.safe diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/driver.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/driver.rs index 41a224c37..782dfd196 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/driver.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/driver.rs @@ -243,6 +243,8 @@ pub(crate) fn test_catalog_with_tma(repo_root: &Path) -> Result { let admission = TmaAdmission { llvm_evidence_profile: "llvm-tma-test".into(), libnvvm_evidence_profile: "libnvvm-tma-test".into(), + cta_llvm_evidence_profile: Some("llvm-tma-test".into()), + cta_libnvvm_evidence_profile: Some("libnvvm-tma-test".into()), reduce_llvm_evidence_profile: Some("llvm-tma-reduce-test".into()), reduce_libnvvm_evidence_profile: Some("libnvvm-tma-reduce-test".into()), runtime_validation: RuntimeValidation::Unexecuted, diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/families/tma.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/families/tma.rs index 99c0ca90a..c0172af85 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/families/tma.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/families/tma.rs @@ -19,14 +19,19 @@ use crate::resolve::guards::*; pub(in crate::resolve) const TMA_BLACKWELL_TARGETS: &str = "sm_100a|sm_101a|sm_103a|sm_110a"; pub(in crate::resolve) const TENSOR_MAP_REPLACE_TARGETS: &str = "sm_100a|sm_100f|sm_103a|sm_103f|sm_110a|sm_110f|sm_120a|sm_120f|sm_121a|sm_121f|sm_90a"; -pub(in crate::resolve) const TMA_OPERATIONS: [TmaOperation; 47] = [ +pub(in crate::resolve) const TMA_OPERATIONS: [TmaOperation; 52] = [ TmaOperation::G2sTile1d, + TmaOperation::G2sCtaTile1d, TmaOperation::G2sTile2d, + TmaOperation::G2sCtaTile2d, TmaOperation::G2sTile2dMulticast, TmaOperation::G2sTile2dMulticastCg2, TmaOperation::G2sTile3d, + TmaOperation::G2sCtaTile3d, TmaOperation::G2sTile4d, + TmaOperation::G2sCtaTile4d, TmaOperation::G2sTile5d, + TmaOperation::G2sCtaTile5d, TmaOperation::S2gTile1d, TmaOperation::S2gTile2d, TmaOperation::S2gTile3d, @@ -409,6 +414,15 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { "CpAsyncBulkTensorG2sTile1dOp", "nvvm.cp_async_bulk_tensor_g2s_tile_1d", ), + TmaOperation::G2sCtaTile1d => ( + "i1034", + "cp_async_bulk_tensor_1d_g2s_cta", + "memory.copy.async.bulk.tensor.g2s.cta.tile.1d", + "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_1d", + "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.1d", + "CpAsyncBulkTensorG2sCtaTile1dOp", + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_1d", + ), TmaOperation::G2sTile2d => ( "i0329", "cp_async_bulk_tensor_2d_g2s", @@ -418,6 +432,15 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { "CpAsyncBulkTensorG2sTile2dOp", "nvvm.cp_async_bulk_tensor_g2s_tile_2d", ), + TmaOperation::G2sCtaTile2d => ( + "i1035", + "cp_async_bulk_tensor_2d_g2s_cta", + "memory.copy.async.bulk.tensor.g2s.cta.tile.2d", + "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_2d", + "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d", + "CpAsyncBulkTensorG2sCtaTile2dOp", + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_2d", + ), TmaOperation::G2sTile2dMulticast => ( "i0330", "cp_async_bulk_tensor_2d_g2s_multicast", @@ -445,6 +468,15 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { "CpAsyncBulkTensorG2sTile3dOp", "nvvm.cp_async_bulk_tensor_g2s_tile_3d", ), + TmaOperation::G2sCtaTile3d => ( + "i1036", + "cp_async_bulk_tensor_3d_g2s_cta", + "memory.copy.async.bulk.tensor.g2s.cta.tile.3d", + "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_3d", + "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.3d", + "CpAsyncBulkTensorG2sCtaTile3dOp", + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_3d", + ), TmaOperation::G2sTile4d => ( "i0333", "cp_async_bulk_tensor_4d_g2s", @@ -454,6 +486,15 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { "CpAsyncBulkTensorG2sTile4dOp", "nvvm.cp_async_bulk_tensor_g2s_tile_4d", ), + TmaOperation::G2sCtaTile4d => ( + "i1037", + "cp_async_bulk_tensor_4d_g2s_cta", + "memory.copy.async.bulk.tensor.g2s.cta.tile.4d", + "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_4d", + "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.4d", + "CpAsyncBulkTensorG2sCtaTile4dOp", + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_4d", + ), TmaOperation::G2sTile5d => ( "i0334", "cp_async_bulk_tensor_5d_g2s", @@ -463,6 +504,15 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { "CpAsyncBulkTensorG2sTile5dOp", "nvvm.cp_async_bulk_tensor_g2s_tile_5d", ), + TmaOperation::G2sCtaTile5d => ( + "i1038", + "cp_async_bulk_tensor_5d_g2s_cta", + "memory.copy.async.bulk.tensor.g2s.cta.tile.5d", + "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_5d", + "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.5d", + "CpAsyncBulkTensorG2sCtaTile5dOp", + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_5d", + ), TmaOperation::S2gTile1d => ( "i0335", "cp_async_bulk_tensor_1d_s2g", @@ -829,12 +879,17 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { let is_g2s = matches!( operation, TmaOperation::G2sTile1d + | TmaOperation::G2sCtaTile1d | TmaOperation::G2sTile2d + | TmaOperation::G2sCtaTile2d | TmaOperation::G2sTile2dMulticast | TmaOperation::G2sTile2dMulticastCg2 | TmaOperation::G2sTile3d + | TmaOperation::G2sCtaTile3d | TmaOperation::G2sTile4d + | TmaOperation::G2sCtaTile4d | TmaOperation::G2sTile5d + | TmaOperation::G2sCtaTile5d ); let is_s2g = matches!( operation, @@ -848,6 +903,7 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { operation, TmaOperation::G2sTile2dMulticast | TmaOperation::G2sTile2dMulticastCg2 ); + let cta_g2s = operation.is_cta_g2s(); let cg2 = operation == TmaOperation::G2sTile2dMulticastCg2; let prefetch_coordinates = operation.prefetch_coordinate_count(); let is_release_fence = matches!( @@ -877,10 +933,19 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { } dialect_operands.extend(["ptr", "ptr", "ptr"]); dialect_operands.extend(std::iter::repeat_n("i32", dimensions.unwrap())); - dialect_operands.extend(["i16", "i64"]); - llvm_arguments.extend(["shared_cluster_ptr", "shared_ptr", "ptr"]); - llvm_arguments.extend(std::iter::repeat_n("i32", dimensions.unwrap())); - llvm_arguments.extend(["i16", "i64", "i1", "i1", "i32"]); + if !cta_g2s { + dialect_operands.push("i16"); + } + dialect_operands.push("i64"); + if cta_g2s { + llvm_arguments.extend(["shared_ptr", "shared_ptr", "ptr"]); + llvm_arguments.extend(std::iter::repeat_n("i32", dimensions.unwrap())); + llvm_arguments.extend(["i64", "i1"]); + } else { + llvm_arguments.extend(["shared_cluster_ptr", "shared_ptr", "ptr"]); + llvm_arguments.extend(std::iter::repeat_n("i32", dimensions.unwrap())); + llvm_arguments.extend(["i16", "i64", "i1", "i1", "i32"]); + } ( if multicast { TmaAdapter::G2sPointersCoordinatesBarrierMaskInjectDefaults @@ -892,6 +957,8 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { true, if multicast { "Starts a multicast TMA tile copy from global to cluster shared memory." + } else if cta_g2s { + "Starts a TMA tile copy from global to the issuing CTA's shared memory." } else { "Starts a TMA tile copy from global to cluster shared memory." }, @@ -1094,7 +1161,11 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { "bulk".into(), "tensor".into(), format!("{}d", dimensions.unwrap()), - "shared::cluster".into(), + if cta_g2s { + "shared::cta".into() + } else { + "shared::cluster".into() + }, "global".into(), "tile".into(), "mbarrier::complete_tx::bytes".into(), @@ -1308,13 +1379,14 @@ pub(in crate::resolve) fn tma_recipe(operation: TmaOperation) -> TmaRecipe { unreachable!("TMA operation category was matched") }; - let minimum_ptx = if blackwell_tma || operation == TmaOperation::ReplaceSwizzleAtomicity { - "8.6" - } else if is_replace || is_fence { - "8.3" - } else { - "8.0" - }; + let minimum_ptx = + if blackwell_tma || cta_g2s || operation == TmaOperation::ReplaceSwizzleAtomicity { + "8.6" + } else if is_replace || is_fence { + "8.3" + } else { + "8.0" + }; let (minimum_sm, targets) = if blackwell_tma { (None, TMA_BLACKWELL_TARGETS) } else if operation == TmaOperation::ReplaceSwizzleAtomicity { @@ -1389,14 +1461,32 @@ pub(in crate::resolve) fn expand_tma_admission( && !admission.libnvvm_evidence_profile.trim().is_empty(), "compact TMA admission requires both backend evidence profiles" ); + let admits_cta_g2s = admission.admits_cta_g2s(); + if admits_cta_g2s { + ensure!( + admission + .cta_llvm_evidence_profile + .as_deref() + .is_some_and(|profile| !profile.trim().is_empty()) + && admission + .cta_libnvvm_evidence_profile + .as_deref() + .is_some_and(|profile| !profile.trim().is_empty()), + "compact TMA admission with CTA-local G2S variants requires cta_llvm_evidence_profile and cta_libnvvm_evidence_profile" + ); + } + let expected_operations = TMA_OPERATIONS + .into_iter() + .filter(|operation| admits_cta_g2s || !operation.is_cta_g2s()) + .collect::>(); ensure!( admission .variants .iter() .map(|variant| variant.operation) - .eq(TMA_OPERATIONS), + .eq(expected_operations.iter().copied()), "compact TMA admission must list all {} operations in canonical order", - TMA_OPERATIONS.len() + expected_operations.len() ); let mut records = admission @@ -1410,6 +1500,22 @@ pub(in crate::resolve) fn expand_tma_admission( recipe.id, recipe.abi_id ); + let llvm_evidence_profile = if recipe.operation.is_cta_g2s() { + admission + .cta_llvm_evidence_profile + .as_ref() + .expect("validated TMA CTA-local LLVM evidence profile") + } else { + &admission.llvm_evidence_profile + }; + let libnvvm_evidence_profile = if recipe.operation.is_cta_g2s() { + admission + .cta_libnvvm_evidence_profile + .as_ref() + .expect("validated TMA CTA-local libNVVM evidence profile") + } else { + &admission.libnvvm_evidence_profile + }; Ok(OverlayIntrinsic { id: recipe.id.into(), abi_id: variant.abi_id.clone(), @@ -1462,7 +1568,7 @@ pub(in crate::resolve) fn expand_tma_admission( OverlayBackendLowering { backend: IntrinsicBackend::LlvmNvptx, mechanism: recipe.llvm_mechanism, - evidence_profile: admission.llvm_evidence_profile.clone(), + evidence_profile: llvm_evidence_profile.clone(), targets: None, minimum_ptx: Some(recipe.minimum_ptx.into()), minimum_sm: recipe.minimum_sm.map(Into::into), @@ -1470,7 +1576,7 @@ pub(in crate::resolve) fn expand_tma_admission( OverlayBackendLowering { backend: IntrinsicBackend::LibNvvm, mechanism: BackendLoweringMechanism::InlinePtx, - evidence_profile: admission.libnvvm_evidence_profile.clone(), + evidence_profile: libnvvm_evidence_profile.clone(), targets: None, minimum_ptx: Some(recipe.minimum_ptx.into()), minimum_sm: recipe.minimum_sm.map(Into::into), @@ -1781,6 +1887,14 @@ pub(in crate::resolve) fn validate_tma_reduction_policy( pub(in crate::resolve) fn tma_imported_properties(operation: TmaOperation) -> Vec { assert_ne!(operation, TmaOperation::Reduce); let dimensions = operation.dimensions(); + if operation.is_cta_g2s() { + return vec![ + format!("ImmArg", dimensions.unwrap() + 4), + "IntrConvergent".into(), + "ReadOnly".into(), + "WriteOnly".into(), + ]; + } if matches!( operation, TmaOperation::G2sTile1d diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/guards.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/guards.rs index 4575d19c8..ac42753c7 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/guards.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/guards.rs @@ -307,10 +307,18 @@ pub(super) fn selection_matches_tma_policy( "TMA_G2S_TILE_CG0_1D", "cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0}}], [$mbar];", ), + TmaOperation::G2sCtaTile1d => ( + "TMA_G2S_CTA_TILE_1D", + "cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0}}], [$mbar];", + ), TmaOperation::G2sTile2d => ( "TMA_G2S_TILE_CG0_2D", "cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1}}], [$mbar];", ), + TmaOperation::G2sCtaTile2d => ( + "TMA_G2S_CTA_TILE_2D", + "cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1}}], [$mbar];", + ), TmaOperation::G2sTile2dMulticast => ( "TMA_G2S_TILE_CG0_2D_MC", "cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster$cg [$dst], [$tmap, {{$d0, $d1}}], [$mbar], $mc;", @@ -323,14 +331,26 @@ pub(super) fn selection_matches_tma_policy( "TMA_G2S_TILE_CG0_3D", "cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2}}], [$mbar];", ), + TmaOperation::G2sCtaTile3d => ( + "TMA_G2S_CTA_TILE_3D", + "cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2}}], [$mbar];", + ), TmaOperation::G2sTile4d => ( "TMA_G2S_TILE_CG0_4D", "cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2, $d3}}], [$mbar];", ), + TmaOperation::G2sCtaTile4d => ( + "TMA_G2S_CTA_TILE_4D", + "cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2, $d3}}], [$mbar];", + ), TmaOperation::G2sTile5d => ( "TMA_G2S_TILE_CG0_5D", "cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2, $d3, $d4}}], [$mbar];", ), + TmaOperation::G2sCtaTile5d => ( + "TMA_G2S_CTA_TILE_5D", + "cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2, $d3, $d4}}], [$mbar];", + ), TmaOperation::S2gTile1d => ( "TMA_TENSOR_S2G_TILE_1D", "cp.async.bulk.tensor.1d.global.shared::cta.tile.bulk_group [$tmap, {{$d0}}], [$src];", @@ -456,29 +476,48 @@ pub(super) fn selection_matches_tma_policy( if matches!( operation, TmaOperation::G2sTile1d + | TmaOperation::G2sCtaTile1d | TmaOperation::G2sTile2d + | TmaOperation::G2sCtaTile2d | TmaOperation::G2sTile2dMulticast | TmaOperation::G2sTile2dMulticastCg2 | TmaOperation::G2sTile3d + | TmaOperation::G2sCtaTile3d | TmaOperation::G2sTile4d + | TmaOperation::G2sCtaTile4d | TmaOperation::G2sTile5d + | TmaOperation::G2sCtaTile5d ) { let dimensions = operation.dimensions().unwrap(); - immediate_bindings.push(crate::model::ImportedImmediateBinding { - argument_index: dimensions + 5, - value: if matches!( - operation, - TmaOperation::G2sTile2dMulticast | TmaOperation::G2sTile2dMulticastCg2 - ) { - -1 - } else { - 0 - }, - }); - immediate_bindings.push(crate::model::ImportedImmediateBinding { - argument_index: dimensions + 6, - value: 0, - }); + if matches!( + operation, + TmaOperation::G2sCtaTile1d + | TmaOperation::G2sCtaTile2d + | TmaOperation::G2sCtaTile3d + | TmaOperation::G2sCtaTile4d + | TmaOperation::G2sCtaTile5d + ) { + immediate_bindings.push(crate::model::ImportedImmediateBinding { + argument_index: dimensions + 4, + value: 0, + }); + } else { + immediate_bindings.push(crate::model::ImportedImmediateBinding { + argument_index: dimensions + 5, + value: if matches!( + operation, + TmaOperation::G2sTile2dMulticast | TmaOperation::G2sTile2dMulticastCg2 + ) { + -1 + } else { + 0 + }, + }); + immediate_bindings.push(crate::model::ImportedImmediateBinding { + argument_index: dimensions + 6, + value: 0, + }); + } } else if matches!( operation, TmaOperation::S2gTile1d diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/overlay.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/overlay.rs index 4b6206f09..703332219 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/overlay.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/overlay.rs @@ -43,6 +43,7 @@ pub(super) const CLUSTER_MEMORY_SHARD_SCHEMA: u32 = 39; pub(super) const CLC_SHARD_SCHEMA: u32 = 40; pub(super) const TMA_SHARD_SCHEMA: u32 = 61; pub(super) const TMA_REDUCTION_SHARD_SCHEMA: u32 = 62; +pub(super) const TMA_CTA_SHARD_SCHEMA: u32 = 65; pub(super) const MBARRIER_EXTENDED_SHARD_SCHEMA: u32 = 40; pub(super) const WGMMA_CONTROL_SHARD_SCHEMA: u32 = 38; pub(super) const TCGEN05_SHARD_SCHEMA: u32 = 42; @@ -571,6 +572,15 @@ pub(super) fn validate_overlay_shard_schema_with_max( "compact TMA reduction admission requires overlay shard schema {}", TMA_REDUCTION_SHARD_SCHEMA ); + ensure!( + shard + .tma + .as_ref() + .is_none_or(|admission| !admission.admits_cta_g2s()) + || shard.schema >= TMA_CTA_SHARD_SCHEMA, + "compact TMA CTA-local G2S admission requires overlay shard schema {}", + TMA_CTA_SHARD_SCHEMA + ); ensure!( shard.mbarrier_extended.is_none() || shard.schema >= MBARRIER_EXTENDED_SHARD_SCHEMA, "compact extended-mbarrier admission requires overlay shard schema {}", diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/fixtures.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/fixtures.rs index 4642f19ee..d4710b663 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/fixtures.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/fixtures.rs @@ -1535,6 +1535,8 @@ pub(super) fn test_tma_admission() -> TmaAdmission { TmaAdmission { llvm_evidence_profile: "llvm-tma-test".into(), libnvvm_evidence_profile: "libnvvm-tma-test".into(), + cta_llvm_evidence_profile: Some("llvm-tma-test".into()), + cta_libnvvm_evidence_profile: Some("libnvvm-tma-test".into()), reduce_llvm_evidence_profile: Some("llvm-tma-reduce-test".into()), reduce_libnvvm_evidence_profile: Some("libnvvm-tma-reduce-test".into()), runtime_validation: RuntimeValidation::Unexecuted, diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/overlay_schema.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/overlay_schema.rs index d10dccace..8be4c9578 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/overlay_schema.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/overlay_schema.rs @@ -49,7 +49,7 @@ fn overlay_manifest_loads_sorted_family_shards() { read_overlay(&repo_root, &repo_root.join("intrinsics/overlay.toml")).unwrap(); assert_eq!(overlay.schema, OVERLAY_SCHEMA); assert_eq!(overlay.shards.len(), 67); - assert_eq!(overlay.intrinsics.len(), 1033); + assert_eq!(overlay.intrinsics.len(), 1038); assert_eq!( overlay .intrinsics @@ -344,6 +344,17 @@ fn overlay_shard_schema_range_is_composable_and_new_fields_fail_closed() { .to_string() .contains("requires overlay shard schema 47") ); + let mut cta_tma = shard(TMA_CTA_SHARD_SCHEMA, None, None); + cta_tma.family = "tma".into(); + cta_tma.tma = Some(test_tma_admission()); + validate_overlay_shard_schema(&cta_tma, path).unwrap(); + cta_tma.schema -= 1; + assert!( + validate_overlay_shard_schema(&cta_tma, path) + .unwrap_err() + .to_string() + .contains("compact TMA CTA-local G2S admission requires overlay shard schema 65") + ); let mut sparse_f16_mma = shard(SPARSE_MMA_F8F6F4_F16_SHARD_SCHEMA, None, None); sparse_f16_mma.sparse_mma_f8f6f4_f16 = Some(test_sparse_mma_f8f6f4_f16_admission()); diff --git a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/tma_clc_cluster.rs b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/tma_clc_cluster.rs index 01f1e00f4..3aa380c48 100644 --- a/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/tma_clc_cluster.rs +++ b/cuda-oxide/crates/cuda-intrinsics-gen/src/resolve/tests/tma_clc_cluster.rs @@ -191,12 +191,17 @@ fn compact_tma_admission_matches_llvm_and_fails_closed() { .collect::>(), [ ("i0328", "cp_async_bulk_tensor_1d_g2s"), + ("i1034", "cp_async_bulk_tensor_1d_g2s_cta"), ("i0329", "cp_async_bulk_tensor_2d_g2s"), + ("i1035", "cp_async_bulk_tensor_2d_g2s_cta"), ("i0330", "cp_async_bulk_tensor_2d_g2s_multicast"), ("i0331", "cp_async_bulk_tensor_2d_g2s_multicast_cg2"), ("i0332", "cp_async_bulk_tensor_3d_g2s"), + ("i1036", "cp_async_bulk_tensor_3d_g2s_cta"), ("i0333", "cp_async_bulk_tensor_4d_g2s"), + ("i1037", "cp_async_bulk_tensor_4d_g2s_cta"), ("i0334", "cp_async_bulk_tensor_5d_g2s"), + ("i1038", "cp_async_bulk_tensor_5d_g2s_cta"), ("i0335", "cp_async_bulk_tensor_1d_s2g"), ("i0336", "cp_async_bulk_tensor_2d_s2g"), ("i0337", "cp_async_bulk_tensor_3d_s2g"), @@ -278,7 +283,7 @@ fn compact_tma_admission_matches_llvm_and_fails_closed() { } assert_eq!( - parse_hardware_target(&records[3]).unwrap(), + parse_hardware_target(&records[5]).unwrap(), CatalogHardwareTarget::AnyOf { alternatives: vec![ CatalogHardwareAlternative::ExactArchitecture { sm: 100 }, @@ -318,6 +323,47 @@ fn compact_tma_admission_matches_llvm_and_fails_closed() { executed.runtime_validation = RuntimeValidation::Executed; assert!(expand_tma_admission(&executed).is_err()); + let mut cta_without_profiles = test_tma_admission(); + cta_without_profiles.cta_llvm_evidence_profile = None; + assert!( + expand_tma_admission(&cta_without_profiles) + .unwrap_err() + .to_string() + .contains("requires cta_llvm_evidence_profile and cta_libnvvm_evidence_profile") + ); + + let cta_count = TMA_OPERATIONS + .iter() + .filter(|operation| operation.is_cta_g2s()) + .count(); + let mut cluster_only = test_tma_admission(); + cluster_only + .variants + .retain(|variant| !variant.operation.is_cta_g2s()); + cluster_only.cta_llvm_evidence_profile = None; + cluster_only.cta_libnvvm_evidence_profile = None; + let cluster_only_records = expand_tma_admission(&cluster_only).unwrap(); + assert_eq!( + cluster_only_records.len(), + TMA_OPERATIONS.len() - cta_count + tma_reduction_matrix().len() + ); + assert!( + cluster_only_records + .iter() + .all(|record| !record.id.ends_with("_g2s_cta")) + ); + + let mut partial_cta = test_tma_admission(); + partial_cta.variants.retain(|variant| { + !variant.operation.is_cta_g2s() || variant.operation == TmaOperation::G2sCtaTile2d + }); + assert!( + expand_tma_admission(&partial_cta) + .unwrap_err() + .to_string() + .contains("canonical order") + ); + let declaration = declarations[records[0].source_record.as_deref().unwrap()]; let mut wrong_adapter = records[0].clone(); wrong_adapter.tma.as_mut().unwrap().adapter = TmaAdapter::NoOperands; @@ -326,11 +372,18 @@ fn compact_tma_admission_matches_llvm_and_fails_closed() { #[test] fn tma_compact_schema_is_reserved_for_aggregation() { - let shard = |schema: u32, include_reductions: bool| { + let shard = |schema: u32, include_reductions: bool, include_cta: bool| { let mut admission = test_tma_admission(); if !include_reductions { admission.reduce_variants.clear(); } + if !include_cta { + admission + .variants + .retain(|variant| !variant.operation.is_cta_g2s()); + admission.cta_llvm_evidence_profile = None; + admission.cta_libnvvm_evidence_profile = None; + } OverlayShardFile { schema, family: "tma".into(), @@ -373,14 +426,14 @@ fn tma_compact_schema_is_reserved_for_aggregation() { let path = Path::new("intrinsics/overlay/tma.toml"); validate_overlay_shard_schema_with_max( - &shard(TMA_SHARD_SCHEMA, false), + &shard(TMA_SHARD_SCHEMA, false, false), path, TMA_REDUCTION_SHARD_SCHEMA, ) .unwrap(); assert!( validate_overlay_shard_schema_with_max( - &shard(TMA_SHARD_SCHEMA - 1, false), + &shard(TMA_SHARD_SCHEMA - 1, false, false), path, TMA_REDUCTION_SHARD_SCHEMA, ) @@ -390,14 +443,14 @@ fn tma_compact_schema_is_reserved_for_aggregation() { ); validate_overlay_shard_schema_with_max( - &shard(TMA_REDUCTION_SHARD_SCHEMA, true), + &shard(TMA_REDUCTION_SHARD_SCHEMA, true, false), path, TMA_REDUCTION_SHARD_SCHEMA, ) .unwrap(); assert!( validate_overlay_shard_schema_with_max( - &shard(TMA_REDUCTION_SHARD_SCHEMA - 1, true), + &shard(TMA_REDUCTION_SHARD_SCHEMA - 1, true, false), path, TMA_REDUCTION_SHARD_SCHEMA, ) @@ -405,6 +458,23 @@ fn tma_compact_schema_is_reserved_for_aggregation() { .to_string() .contains("compact TMA reduction admission requires overlay shard schema 62") ); + + validate_overlay_shard_schema_with_max( + &shard(TMA_CTA_SHARD_SCHEMA, true, true), + path, + TMA_CTA_SHARD_SCHEMA, + ) + .unwrap(); + assert!( + validate_overlay_shard_schema_with_max( + &shard(TMA_CTA_SHARD_SCHEMA - 1, true, true), + path, + TMA_CTA_SHARD_SCHEMA, + ) + .unwrap_err() + .to_string() + .contains("compact TMA CTA-local G2S admission requires overlay shard schema 65") + ); } #[test] diff --git a/cuda-oxide/crates/cuda-intrinsics/src/generated/abi_v1.rs b/cuda-oxide/crates/cuda-intrinsics/src/generated/abi_v1.rs index 698a64daf..f657dd466 100644 --- a/cuda-oxide/crates/cuda-intrinsics/src/generated/abi_v1.rs +++ b/cuda-oxide/crates/cuda-intrinsics/src/generated/abi_v1.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Raw ABI functions recognized by cuda-oxide. @@ -1181,6 +1181,22 @@ pub unsafe fn i0328(_arg0: *mut u8, _arg1: *const u8, _arg2: i32, _arg3: *mut u6 ) } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// Catalog ID: `cp_async_bulk_tensor_1d_g2s_cta`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.1d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_1d`; expects PTX `cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];`. +/// Available on `sm_90+` targets from PTX 8.6. +/// +/// # Safety +/// The destination and barrier must lie in the issuing CTA's own shared memory, and the tensor map must be a live descriptor for this dimensionality. +/// A generic destination or barrier pointer is converted with `cvta.to.shared`, so a cluster-shared pointer from `map_shared_rank` would yield a wrong local offset; copies into another CTA's shared memory use the cluster form. +/// Keep every object alive until the asynchronous copy completes. Only the designated issuing thread may start this transfer. +#[inline(never)] +pub unsafe fn i1034(_arg0: *mut u8, _arg1: *const u8, _arg2: i32, _arg3: *mut u64) -> () { + unreachable!( + "generated CUDA intrinsic `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034` executed outside device compilation" + ) +} + /// Starts a TMA tile copy from shared to global memory. /// /// Catalog ID: `cp_async_bulk_tensor_1d_s2g`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.1d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_1d`; expects PTX `cp.async.bulk.tensor.1d.global.shared::cta.tile.bulk_group [
], [
];`. @@ -1217,6 +1233,28 @@ pub unsafe fn i0329( ) } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// Catalog ID: `cp_async_bulk_tensor_2d_g2s_cta`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_2d`; expects PTX `cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];`. +/// Available on `sm_90+` targets from PTX 8.6. +/// +/// # Safety +/// The destination and barrier must lie in the issuing CTA's own shared memory, and the tensor map must be a live descriptor for this dimensionality. +/// A generic destination or barrier pointer is converted with `cvta.to.shared`, so a cluster-shared pointer from `map_shared_rank` would yield a wrong local offset; copies into another CTA's shared memory use the cluster form. +/// Keep every object alive until the asynchronous copy completes. Only the designated issuing thread may start this transfer. +#[inline(never)] +pub unsafe fn i1035( + _arg0: *mut u8, + _arg1: *const u8, + _arg2: i32, + _arg3: i32, + _arg4: *mut u64, +) -> () { + unreachable!( + "generated CUDA intrinsic `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035` executed outside device compilation" + ) +} + /// Starts a multicast TMA tile copy from global to cluster shared memory. /// /// Catalog ID: `cp_async_bulk_tensor_2d_g2s_multicast`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d` from `int_nvvm_cp_async_bulk_tensor_g2s_tile_2d`; expects PTX `cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [
], [
], [
], ;`. @@ -1298,6 +1336,29 @@ pub unsafe fn i0332( ) } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// Catalog ID: `cp_async_bulk_tensor_3d_g2s_cta`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.3d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_3d`; expects PTX `cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];`. +/// Available on `sm_90+` targets from PTX 8.6. +/// +/// # Safety +/// The destination and barrier must lie in the issuing CTA's own shared memory, and the tensor map must be a live descriptor for this dimensionality. +/// A generic destination or barrier pointer is converted with `cvta.to.shared`, so a cluster-shared pointer from `map_shared_rank` would yield a wrong local offset; copies into another CTA's shared memory use the cluster form. +/// Keep every object alive until the asynchronous copy completes. Only the designated issuing thread may start this transfer. +#[inline(never)] +pub unsafe fn i1036( + _arg0: *mut u8, + _arg1: *const u8, + _arg2: i32, + _arg3: i32, + _arg4: i32, + _arg5: *mut u64, +) -> () { + unreachable!( + "generated CUDA intrinsic `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036` executed outside device compilation" + ) +} + /// Starts a TMA tile copy from shared to global memory. /// /// Catalog ID: `cp_async_bulk_tensor_3d_s2g`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.3d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_3d`; expects PTX `cp.async.bulk.tensor.3d.global.shared::cta.tile.bulk_group [
], [
];`. @@ -1336,6 +1397,30 @@ pub unsafe fn i0333( ) } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// Catalog ID: `cp_async_bulk_tensor_4d_g2s_cta`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.4d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_4d`; expects PTX `cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];`. +/// Available on `sm_90+` targets from PTX 8.6. +/// +/// # Safety +/// The destination and barrier must lie in the issuing CTA's own shared memory, and the tensor map must be a live descriptor for this dimensionality. +/// A generic destination or barrier pointer is converted with `cvta.to.shared`, so a cluster-shared pointer from `map_shared_rank` would yield a wrong local offset; copies into another CTA's shared memory use the cluster form. +/// Keep every object alive until the asynchronous copy completes. Only the designated issuing thread may start this transfer. +#[inline(never)] +pub unsafe fn i1037( + _arg0: *mut u8, + _arg1: *const u8, + _arg2: i32, + _arg3: i32, + _arg4: i32, + _arg5: i32, + _arg6: *mut u64, +) -> () { + unreachable!( + "generated CUDA intrinsic `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037` executed outside device compilation" + ) +} + /// Starts a TMA tile copy from shared to global memory. /// /// Catalog ID: `cp_async_bulk_tensor_4d_s2g`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.4d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_4d`; expects PTX `cp.async.bulk.tensor.4d.global.shared::cta.tile.bulk_group [
], [
];`. @@ -1383,6 +1468,32 @@ pub unsafe fn i0334( ) } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +/// +/// Catalog ID: `cp_async_bulk_tensor_5d_g2s_cta`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.5d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_5d`; expects PTX `cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];`. +/// Available on `sm_90+` targets from PTX 8.6. +/// +/// # Safety +/// The destination and barrier must lie in the issuing CTA's own shared memory, and the tensor map must be a live descriptor for this dimensionality. +/// A generic destination or barrier pointer is converted with `cvta.to.shared`, so a cluster-shared pointer from `map_shared_rank` would yield a wrong local offset; copies into another CTA's shared memory use the cluster form. +/// Keep every object alive until the asynchronous copy completes. Only the designated issuing thread may start this transfer. +#[allow(clippy::too_many_arguments)] +#[inline(never)] +pub unsafe fn i1038( + _arg0: *mut u8, + _arg1: *const u8, + _arg2: i32, + _arg3: i32, + _arg4: i32, + _arg5: i32, + _arg6: i32, + _arg7: *mut u64, +) -> () { + unreachable!( + "generated CUDA intrinsic `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038` executed outside device compilation" + ) +} + /// Starts a TMA tile copy from shared to global memory. /// /// Catalog ID: `cp_async_bulk_tensor_5d_s2g`. Source: LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.5d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_5d`; expects PTX `cp.async.bulk.tensor.5d.global.shared::cta.tile.bulk_group [
], [
];`. diff --git a/cuda-oxide/crates/cuda-intrinsics/src/generated/mod.rs b/cuda-oxide/crates/cuda-intrinsics/src/generated/mod.rs index 814cc0e02..c6488d7ac 100644 --- a/cuda-oxide/crates/cuda-intrinsics/src/generated/mod.rs +++ b/cuda-oxide/crates/cuda-intrinsics/src/generated/mod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. #[doc(hidden)] @@ -1098,6 +1098,11 @@ pub mod tma { pub use crate::__cuda_oxide_intrinsic_abi_v1::i0984 as cp_async_bulk_tensor_reduce_xor_im2col_3d; pub use crate::__cuda_oxide_intrinsic_abi_v1::i0985 as cp_async_bulk_tensor_reduce_xor_im2col_4d; pub use crate::__cuda_oxide_intrinsic_abi_v1::i0986 as cp_async_bulk_tensor_reduce_xor_im2col_5d; + pub use crate::__cuda_oxide_intrinsic_abi_v1::i1034 as cp_async_bulk_tensor_1d_g2s_cta; + pub use crate::__cuda_oxide_intrinsic_abi_v1::i1035 as cp_async_bulk_tensor_2d_g2s_cta; + pub use crate::__cuda_oxide_intrinsic_abi_v1::i1036 as cp_async_bulk_tensor_3d_g2s_cta; + pub use crate::__cuda_oxide_intrinsic_abi_v1::i1037 as cp_async_bulk_tensor_4d_g2s_cta; + pub use crate::__cuda_oxide_intrinsic_abi_v1::i1038 as cp_async_bulk_tensor_5d_g2s_cta; } /// Generated `warp` intrinsic source API. @@ -1694,6 +1699,14 @@ mod tests { assert_eq!(public as usize, canonical as usize); } #[test] + fn public_cp_async_bulk_tensor_1d_g2s_cta_reexports_abi_i1034() { + let public: unsafe fn(*mut u8, *const u8, i32, *mut u64) -> () = + super::tma::cp_async_bulk_tensor_1d_g2s_cta; + let canonical: unsafe fn(*mut u8, *const u8, i32, *mut u64) -> () = + super::__cuda_oxide_intrinsic_abi_v1::i1034; + assert_eq!(public as usize, canonical as usize); + } + #[test] fn public_cp_async_bulk_tensor_1d_s2g_reexports_abi_i0335() { let public: unsafe fn(*const u8, *const u8, i32) -> () = super::tma::cp_async_bulk_tensor_1d_s2g; @@ -1710,6 +1723,14 @@ mod tests { assert_eq!(public as usize, canonical as usize); } #[test] + fn public_cp_async_bulk_tensor_2d_g2s_cta_reexports_abi_i1035() { + let public: unsafe fn(*mut u8, *const u8, i32, i32, *mut u64) -> () = + super::tma::cp_async_bulk_tensor_2d_g2s_cta; + let canonical: unsafe fn(*mut u8, *const u8, i32, i32, *mut u64) -> () = + super::__cuda_oxide_intrinsic_abi_v1::i1035; + assert_eq!(public as usize, canonical as usize); + } + #[test] fn public_cp_async_bulk_tensor_2d_g2s_multicast_reexports_abi_i0330() { let public: unsafe fn(*mut u8, *const u8, i32, i32, *mut u64, u16) -> () = super::tma::cp_async_bulk_tensor_2d_g2s_multicast; @@ -1742,6 +1763,14 @@ mod tests { assert_eq!(public as usize, canonical as usize); } #[test] + fn public_cp_async_bulk_tensor_3d_g2s_cta_reexports_abi_i1036() { + let public: unsafe fn(*mut u8, *const u8, i32, i32, i32, *mut u64) -> () = + super::tma::cp_async_bulk_tensor_3d_g2s_cta; + let canonical: unsafe fn(*mut u8, *const u8, i32, i32, i32, *mut u64) -> () = + super::__cuda_oxide_intrinsic_abi_v1::i1036; + assert_eq!(public as usize, canonical as usize); + } + #[test] fn public_cp_async_bulk_tensor_3d_s2g_reexports_abi_i0337() { let public: unsafe fn(*const u8, *const u8, i32, i32, i32) -> () = super::tma::cp_async_bulk_tensor_3d_s2g; @@ -1758,6 +1787,14 @@ mod tests { assert_eq!(public as usize, canonical as usize); } #[test] + fn public_cp_async_bulk_tensor_4d_g2s_cta_reexports_abi_i1037() { + let public: unsafe fn(*mut u8, *const u8, i32, i32, i32, i32, *mut u64) -> () = + super::tma::cp_async_bulk_tensor_4d_g2s_cta; + let canonical: unsafe fn(*mut u8, *const u8, i32, i32, i32, i32, *mut u64) -> () = + super::__cuda_oxide_intrinsic_abi_v1::i1037; + assert_eq!(public as usize, canonical as usize); + } + #[test] fn public_cp_async_bulk_tensor_4d_s2g_reexports_abi_i0338() { let public: unsafe fn(*const u8, *const u8, i32, i32, i32, i32) -> () = super::tma::cp_async_bulk_tensor_4d_s2g; @@ -1774,6 +1811,14 @@ mod tests { assert_eq!(public as usize, canonical as usize); } #[test] + fn public_cp_async_bulk_tensor_5d_g2s_cta_reexports_abi_i1038() { + let public: unsafe fn(*mut u8, *const u8, i32, i32, i32, i32, i32, *mut u64) -> () = + super::tma::cp_async_bulk_tensor_5d_g2s_cta; + let canonical: unsafe fn(*mut u8, *const u8, i32, i32, i32, i32, i32, *mut u64) -> () = + super::__cuda_oxide_intrinsic_abi_v1::i1038; + assert_eq!(public as usize, canonical as usize); + } + #[test] fn public_cp_async_bulk_tensor_5d_s2g_reexports_abi_i0339() { let public: unsafe fn(*const u8, *const u8, i32, i32, i32, i32, i32) -> () = super::tma::cp_async_bulk_tensor_5d_s2g; diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/active_mask.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/active_mask.rs index 093d3bf02..762b7fe47 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/active_mask.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/active_mask.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `active_mask` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/clc.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/clc.rs index 688a055fe..fce519434 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/clc.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/clc.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `clc` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cluster_barrier.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cluster_barrier.rs index 2c88800e4..6080a1396 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cluster_barrier.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cluster_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `cluster_barrier` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cluster_memory.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cluster_memory.rs index 49d2527e3..35c3b5a4b 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cluster_memory.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cluster_memory.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `cluster_memory` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cp_async.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cp_async.rs index 7224a18db..ef7573f55 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cp_async.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/cp_async.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `cp_async` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/debug_control.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/debug_control.rs index 48b8883a5..f1c34aea5 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/debug_control.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/debug_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `debug_control` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/dotprod.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/dotprod.rs index f451ca2cd..295b13295 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/dotprod.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/dotprod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `dotprod` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/elect.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/elect.rs index 537884cd4..4b884cc74 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/elect.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/elect.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `elect` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/execution_control.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/execution_control.rs index e1b206d73..8b1784864 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/execution_control.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/execution_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `execution_control` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/extended_minmax.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/extended_minmax.rs index 58981a1e8..c0dbaa2d2 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/extended_minmax.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/extended_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `extended_minmax` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/integer_minmax.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/integer_minmax.rs index 89807769d..988a33ef0 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/integer_minmax.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/integer_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `integer_minmax` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/ldmatrix.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/ldmatrix.rs index 352537c28..1109ff8fa 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/ldmatrix.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/ldmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `ldmatrix` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mbarrier_basic.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mbarrier_basic.rs index c8c5f39b6..2cf4b179a 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mbarrier_basic.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mbarrier_basic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `mbarrier_basic` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mbarrier_extended.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mbarrier_extended.rs index efc2c5af9..cca8493be 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mbarrier_extended.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mbarrier_extended.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `mbarrier_extended` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mod.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mod.rs index a9f16d3a8..06501cfd4 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mod.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/mod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated target requirements and separately imported LLVM/selection facts. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/movmatrix.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/movmatrix.rs index 75742c9f5..21ccb310a 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/movmatrix.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/movmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `movmatrix` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_alu.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_alu.rs index 1d0c819e5..662cd8287 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_alu.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_alu.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `packed_alu` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_atomic.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_atomic.rs index 47a457cd8..9a6180d16 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_atomic.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_atomic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `packed_atomic` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_conversion.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_conversion.rs index d32c5ca86..429b6f3f2 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_conversion.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/packed_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `packed_conversion` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/prmt.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/prmt.rs index 6e2ea5516..b2fb0a2fb 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/prmt.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/prmt.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `prmt` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/redux.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/redux.rs index cb61ea288..15a6571b3 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/redux.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/redux.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `redux` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/register_mma.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/register_mma.rs index 094431531..2d614270c 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/register_mma.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/register_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `register_mma` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_arithmetic.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_arithmetic.rs index 22caa376f..7c011622f 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_arithmetic.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_arithmetic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `scalar_arithmetic` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_conversion.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_conversion.rs index 45b6c0a6c..124621299 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_conversion.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `scalar_conversion` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_math.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_math.rs index 12ee21415..95db81f70 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_math.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/scalar_math.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `scalar_math` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sparse_mma.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sparse_mma.rs index 3fad449c5..77aac0373 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sparse_mma.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sparse_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `sparse_mma` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sreg.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sreg.rs index 702bb761a..dce8c8e25 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sreg.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `sreg` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/stmatrix.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/stmatrix.rs index ee826a919..e006284d6 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/stmatrix.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/stmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `stmatrix` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sync.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sync.rs index 90a5a6f5c..c545d4bc5 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sync.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/sync.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `sync` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_cp.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_cp.rs index 2eaa0883b..05ad33c61 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_cp.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_cp.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `tcgen05_cp` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_ld.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_ld.rs index 3ba5b6bc9..952755650 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_ld.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_ld.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `tcgen05_ld` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_mma.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_mma.rs index 5b33dcaab..c0ea24b29 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_mma.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `tcgen05_mma` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_mma_ws.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_mma_ws.rs index 966f5d125..c996991f7 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_mma_ws.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_mma_ws.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `tcgen05_mma_ws` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_other.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_other.rs index 439101422..91934ba52 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_other.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_other.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `tcgen05_other` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_st.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_st.rs index b26d92081..6b92fbf3d 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_st.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tcgen05_st.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `tcgen05_st` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/active_mask.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/active_mask.rs index a91888a9e..dfb74bc78 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/active_mask.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/active_mask.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `active_mask` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/clc.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/clc.rs index 28ac52b10..4ab2393cd 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/clc.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/clc.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `clc` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cluster_barrier.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cluster_barrier.rs index 8bfd36c97..82ea1f733 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cluster_barrier.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cluster_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `cluster_barrier` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cluster_memory.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cluster_memory.rs index 25eaa798b..075b1b73b 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cluster_memory.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cluster_memory.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `cluster_memory` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cp_async.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cp_async.rs index ebcf09356..534646152 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cp_async.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/cp_async.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `cp_async` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/debug_control.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/debug_control.rs index 0fba32d40..e9eb94805 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/debug_control.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/debug_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `debug_control` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/dotprod.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/dotprod.rs index bde2e5e03..41ba8a614 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/dotprod.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/dotprod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `dotprod` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/elect.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/elect.rs index e76d955ec..b45dec5b8 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/elect.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/elect.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `elect` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/execution_control.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/execution_control.rs index 75ddd78bd..deacdfef7 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/execution_control.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/execution_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `execution_control` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/extended_minmax.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/extended_minmax.rs index 035edab56..6a4498941 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/extended_minmax.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/extended_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `extended_minmax` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/integer_minmax.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/integer_minmax.rs index 8e53cab66..6d66affba 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/integer_minmax.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/integer_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `integer_minmax` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/ldmatrix.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/ldmatrix.rs index 02ebdb7bd..261b4ac11 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/ldmatrix.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/ldmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `ldmatrix` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mbarrier_basic.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mbarrier_basic.rs index 1dec957e1..7b70285b8 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mbarrier_basic.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mbarrier_basic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `mbarrier_basic` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mbarrier_extended.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mbarrier_extended.rs index 6af07e4e1..bb920b133 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mbarrier_extended.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mbarrier_extended.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `mbarrier_extended` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mod.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mod.rs index 8a1c0f797..7508690da 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mod.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/mod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests, grouped per family shard. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/movmatrix.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/movmatrix.rs index b4d35cc17..3a7c7de79 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/movmatrix.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/movmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `movmatrix` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_alu.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_alu.rs index 4f85f04af..ddf741939 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_alu.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_alu.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `packed_alu` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_atomic.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_atomic.rs index 49d50ea99..2ecf439f6 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_atomic.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_atomic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `packed_atomic` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_conversion.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_conversion.rs index e8995a5f9..674c543f5 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_conversion.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/packed_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `packed_conversion` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/prmt.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/prmt.rs index ab4efef6d..7387fa58f 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/prmt.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/prmt.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `prmt` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/redux.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/redux.rs index fdbce3ea5..28a2cbc78 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/redux.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/redux.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `redux` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/register_mma_float.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/register_mma_float.rs index ff14c1f0f..e568072fb 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/register_mma_float.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/register_mma_float.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `register_mma_float` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/register_mma_s32.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/register_mma_s32.rs index a3142f758..715449d0d 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/register_mma_s32.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/register_mma_s32.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `register_mma_s32` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_arithmetic.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_arithmetic.rs index 5a905c322..380646c94 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_arithmetic.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_arithmetic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `scalar_arithmetic` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_conversion.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_conversion.rs index a752dbd24..45cac4183 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_conversion.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `scalar_conversion` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_math.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_math.rs index 30bf7c6ad..4c39a3e6d 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_math.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/scalar_math.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `scalar_math` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sparse_mma.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sparse_mma.rs index 699539ac9..c8650d137 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sparse_mma.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sparse_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `sparse_mma` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sreg.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sreg.rs index 1d3657b88..81e17610b 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sreg.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `sreg` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/stmatrix.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/stmatrix.rs index 948959dca..1e75444e5 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/stmatrix.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/stmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `stmatrix` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sync.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sync.rs index f4730fecb..d0582987c 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sync.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/sync.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `sync` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_cp.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_cp.rs index dc2676f8b..5d32dcefa 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_cp.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_cp.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `tcgen05_cp` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_ld.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_ld.rs index 3e2950492..0e30bd672 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_ld.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_ld.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `tcgen05_ld` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_mma.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_mma.rs index 8ae2ee5ad..96d2c473b 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_mma.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `tcgen05_mma` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_mma_ws.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_mma_ws.rs index 7ed5e540e..28611a174 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_mma_ws.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_mma_ws.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `tcgen05_mma_ws` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_other.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_other.rs index 4c5fc7ce9..91e22ce0d 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_other.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_other.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `tcgen05_other` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_st.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_st.rs index 155020b73..6db2425d6 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_st.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tcgen05_st.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `tcgen05_st` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tma.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tma.rs index 6468154b5..4300e46d3 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tma.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/tma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `tma` intrinsics. @@ -6643,3 +6643,348 @@ fn cp_async_bulk_tensor_reduce_xor_im2col_5d_target_matches_the_catalog() { ] as &[&str] ); } + +#[test] +fn cp_async_bulk_tensor_1d_g2s_cta_target_matches_the_catalog() { + let target = generated_intrinsic_target_by_marker("v1:i1034").unwrap(); + assert_eq!(target.id, "cp_async_bulk_tensor_1d_g2s_cta"); + assert_eq!(target.abi_id, "i1034"); + assert_eq!( + target.dialect_op, + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_1d" + ); + assert_eq!(target.variant, GeneratedIntrinsicVariant::Scalar); + assert_eq!(target.requirement.minimum_ptx.encoded(), 86); + assert_eq!( + target.requirement.hardware, + const { GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm(90)]) } + ); + assert_eq!( + target.backend_requirements, + const { + &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ] + } + ); + assert_eq!( + target.selections, + &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_1D", + asm: "cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 5, + value: 0 + }] + } + },] + ); + assert_eq!( + target.llvm.unwrap().properties, + &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly" + ] as &[&str] + ); +} + +#[test] +fn cp_async_bulk_tensor_2d_g2s_cta_target_matches_the_catalog() { + let target = generated_intrinsic_target_by_marker("v1:i1035").unwrap(); + assert_eq!(target.id, "cp_async_bulk_tensor_2d_g2s_cta"); + assert_eq!(target.abi_id, "i1035"); + assert_eq!( + target.dialect_op, + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_2d" + ); + assert_eq!(target.variant, GeneratedIntrinsicVariant::Scalar); + assert_eq!(target.requirement.minimum_ptx.encoded(), 86); + assert_eq!( + target.requirement.hardware, + const { GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm(90)]) } + ); + assert_eq!( + target.backend_requirements, + const { + &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ] + } + ); + assert_eq!( + target.selections, + &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_2D", + asm: "cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 6, + value: 0 + }] + } + },] + ); + assert_eq!( + target.llvm.unwrap().properties, + &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly" + ] as &[&str] + ); +} + +#[test] +fn cp_async_bulk_tensor_3d_g2s_cta_target_matches_the_catalog() { + let target = generated_intrinsic_target_by_marker("v1:i1036").unwrap(); + assert_eq!(target.id, "cp_async_bulk_tensor_3d_g2s_cta"); + assert_eq!(target.abi_id, "i1036"); + assert_eq!( + target.dialect_op, + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_3d" + ); + assert_eq!(target.variant, GeneratedIntrinsicVariant::Scalar); + assert_eq!(target.requirement.minimum_ptx.encoded(), 86); + assert_eq!( + target.requirement.hardware, + const { GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm(90)]) } + ); + assert_eq!( + target.backend_requirements, + const { + &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ] + } + ); + assert_eq!( + target.selections, + &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_3D", + asm: "cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 7, + value: 0 + }] + } + },] + ); + assert_eq!( + target.llvm.unwrap().properties, + &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly" + ] as &[&str] + ); +} + +#[test] +fn cp_async_bulk_tensor_4d_g2s_cta_target_matches_the_catalog() { + let target = generated_intrinsic_target_by_marker("v1:i1037").unwrap(); + assert_eq!(target.id, "cp_async_bulk_tensor_4d_g2s_cta"); + assert_eq!(target.abi_id, "i1037"); + assert_eq!( + target.dialect_op, + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_4d" + ); + assert_eq!(target.variant, GeneratedIntrinsicVariant::Scalar); + assert_eq!(target.requirement.minimum_ptx.encoded(), 86); + assert_eq!( + target.requirement.hardware, + const { GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm(90)]) } + ); + assert_eq!( + target.backend_requirements, + const { + &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ] + } + ); + assert_eq!( + target.selections, + &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_4D", + asm: "cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2, $d3}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 8, + value: 0 + }] + } + },] + ); + assert_eq!( + target.llvm.unwrap().properties, + &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly" + ] as &[&str] + ); +} + +#[test] +fn cp_async_bulk_tensor_5d_g2s_cta_target_matches_the_catalog() { + let target = generated_intrinsic_target_by_marker("v1:i1038").unwrap(); + assert_eq!(target.id, "cp_async_bulk_tensor_5d_g2s_cta"); + assert_eq!(target.abi_id, "i1038"); + assert_eq!( + target.dialect_op, + "nvvm.cp_async_bulk_tensor_g2s_cta_tile_5d" + ); + assert_eq!(target.variant, GeneratedIntrinsicVariant::Scalar); + assert_eq!(target.requirement.minimum_ptx.encoded(), 86); + assert_eq!( + target.requirement.hardware, + const { GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm(90)]) } + ); + assert_eq!( + target.backend_requirements, + const { + &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ] + } + ); + assert_eq!( + target.selections, + &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_5D", + asm: "cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2, $d3, $d4}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 9, + value: 0 + }] + } + },] + ); + assert_eq!( + target.llvm.unwrap().properties, + &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly" + ] as &[&str] + ); +} diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/vote.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/vote.rs index 7aadebb14..454968edb 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/vote.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/vote.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `vote` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_barrier.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_barrier.rs index d1faf65d4..8bdc1e7cd 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_barrier.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `warp_barrier` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_match.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_match.rs index 404e804d4..a82f03238 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_match.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_match.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `warp_match` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_shuffle.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_shuffle.rs index d69610fcb..f9bda00d6 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_shuffle.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/warp_shuffle.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `warp_shuffle` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/wgmma_control.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/wgmma_control.rs index 86ec25cef..f553a689b 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/wgmma_control.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tests/wgmma_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target tests: `wgmma_control` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tma.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tma.rs index ac7d58f3e..9beadfb3b 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tma.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/tma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `tma` intrinsics. @@ -5670,4 +5670,294 @@ pub(super) const TARGETS: &[GeneratedIntrinsicTarget] = &[ result_range: None, }), }, + GeneratedIntrinsicTarget { + marker: "v1:i1034", + id: "cp_async_bulk_tensor_1d_g2s_cta", + abi_id: "i1034", + dialect_op: "nvvm.cp_async_bulk_tensor_g2s_cta_tile_1d", + variant: GeneratedIntrinsicVariant::Scalar, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm( + 90, + )]), + }, + backend_requirements: &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ], + selections: &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_1D", + asm: "cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86", + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 5, + value: 0, + }], + }, + }], + llvm: Some(GeneratedLlvmFacts { + properties: &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly", + ], + result_no_undef: false, + result_range: None, + }), + }, + GeneratedIntrinsicTarget { + marker: "v1:i1035", + id: "cp_async_bulk_tensor_2d_g2s_cta", + abi_id: "i1035", + dialect_op: "nvvm.cp_async_bulk_tensor_g2s_cta_tile_2d", + variant: GeneratedIntrinsicVariant::Scalar, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm( + 90, + )]), + }, + backend_requirements: &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ], + selections: &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_2D", + asm: "cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86", + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 6, + value: 0, + }], + }, + }], + llvm: Some(GeneratedLlvmFacts { + properties: &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly", + ], + result_no_undef: false, + result_range: None, + }), + }, + GeneratedIntrinsicTarget { + marker: "v1:i1036", + id: "cp_async_bulk_tensor_3d_g2s_cta", + abi_id: "i1036", + dialect_op: "nvvm.cp_async_bulk_tensor_g2s_cta_tile_3d", + variant: GeneratedIntrinsicVariant::Scalar, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm( + 90, + )]), + }, + backend_requirements: &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ], + selections: &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_3D", + asm: "cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86", + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 7, + value: 0, + }], + }, + }], + llvm: Some(GeneratedLlvmFacts { + properties: &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly", + ], + result_no_undef: false, + result_range: None, + }), + }, + GeneratedIntrinsicTarget { + marker: "v1:i1037", + id: "cp_async_bulk_tensor_4d_g2s_cta", + abi_id: "i1037", + dialect_op: "nvvm.cp_async_bulk_tensor_g2s_cta_tile_4d", + variant: GeneratedIntrinsicVariant::Scalar, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm( + 90, + )]), + }, + backend_requirements: &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ], + selections: &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_4D", + asm: "cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2, $d3}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86", + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 8, + value: 0, + }], + }, + }], + llvm: Some(GeneratedLlvmFacts { + properties: &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly", + ], + result_no_undef: false, + result_range: None, + }), + }, + GeneratedIntrinsicTarget { + marker: "v1:i1038", + id: "cp_async_bulk_tensor_5d_g2s_cta", + abi_id: "i1038", + dialect_op: "nvvm.cp_async_bulk_tensor_g2s_cta_tile_5d", + variant: GeneratedIntrinsicVariant::Scalar, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[GeneratedHardwareAlternative::MinimumSm( + 90, + )]), + }, + backend_requirements: &[ + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LlvmNvptx, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + GeneratedBackendRequirement { + backend: GeneratedIntrinsicBackend::LibNvvm, + requirement: GeneratedTargetRequirement { + minimum_ptx: GeneratedPtxVersion::from_encoded(86), + hardware: GeneratedHardwareTarget::AnyOf(&[ + GeneratedHardwareAlternative::MinimumSm(90), + ]), + }, + }, + ], + selections: &[GeneratedSelectionAlternative { + source_record: "TMA_G2S_CTA_TILE_5D", + asm: "cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2, $d3, $d4}}], [$mbar];", + predicates: &[ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86", + ], + constraints: GeneratedSelectionConstraints { + address_space: None, + immediate_bindings: &[GeneratedImmediateBinding { + argument_index: 9, + value: 0, + }], + }, + }], + llvm: Some(GeneratedLlvmFacts { + properties: &[ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly", + ], + result_no_undef: false, + result_range: None, + }), + }, ]; diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/vote.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/vote.rs index 4bfa77552..e83ef298c 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/vote.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/vote.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `vote` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_barrier.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_barrier.rs index cccaf4dbd..cc7a79ffc 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_barrier.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `warp_barrier` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_match.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_match.rs index 000d21b37..02d5833fc 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_match.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_match.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `warp_match` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_shuffle.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_shuffle.rs index 878925151..1f5ded8fb 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_shuffle.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/warp_shuffle.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `warp_shuffle` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/wgmma_control.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/wgmma_control.rs index 30055018e..bdb2b1413 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/wgmma_control.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/generated_intrinsic_targets/wgmma_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated intrinsic target records: `wgmma_control` intrinsics. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/target/detect.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/target/detect.rs index 4e067ece0..20ce82e87 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/target/detect.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/target/detect.rs @@ -574,15 +574,17 @@ pub(super) fn contains_tma_cta_group_features(contents: &str) -> bool { }) } -/// Checks TMA copies whose destination is CTA-local shared memory. +/// Checks TMA copies whose destination is CTA-local shared memory, as inline +/// PTX or as the typed `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.*` intrinsics. /// /// `.shared::cta` already existed as a source state space for shared-to-global /// copies, so the following `.global` source qualifier is part of the match. /// The destination form was introduced in PTX 8.6 but is valid on sm_90. pub(super) fn contains_tma_shared_cta_destination(contents: &str) -> bool { - contents.split(';').any(|statement| { - statement.contains("cp.async.bulk.") && statement.contains(".shared::cta.global") - }) + contents.contains("cp.async.bulk.tensor.g2s.cta.") + || contents.split(';').any(|statement| { + statement.contains("cp.async.bulk.") && statement.contains(".shared::cta.global") + }) } /// Checks PTX 8.6 TMA modifiers with a generic sm_100 architecture floor. diff --git a/cuda-oxide/crates/cuda-oxide-codegen/src/target/tests/detection_async_memory.rs b/cuda-oxide/crates/cuda-oxide-codegen/src/target/tests/detection_async_memory.rs index a81d61588..9c644810b 100644 --- a/cuda-oxide/crates/cuda-oxide-codegen/src/target/tests/detection_async_memory.rs +++ b/cuda-oxide/crates/cuda-oxide-codegen/src/target/tests/detection_async_memory.rs @@ -58,6 +58,15 @@ fn tma_and_wgmma_raise_their_independent_ptx_floors() { assert_eq!(shared_cta_requirements.features, DetectedFeatures::Tma); assert_eq!(shared_cta_requirements.ptx_isa, PtxIsaRequirement::new(86)); + let typed_shared_cta = "call void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d(ptr addrspace(3) %0, ptr addrspace(3) %1, ptr %2, i32 %3, i32 %4, i64 0, i1 false)"; + assert!(contains_tma_shared_cta_destination(typed_shared_cta)); + let typed_requirements = detect_module_requirements_in_llvm_text(typed_shared_cta); + assert_eq!(typed_requirements.features, DetectedFeatures::Tma); + assert_eq!(typed_requirements.ptx_isa, PtxIsaRequirement::new(86)); + assert!(!contains_tma_shared_cta_destination( + "call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d(ptr addrspace(7) %0)" + )); + let shared_source = "cp.async.bulk.tensor.2d.global.shared::cta.tile.bulk_group;"; assert!(!contains_tma_shared_cta_destination(shared_source)); assert_eq!( diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/active_mask.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/active_mask.rs index a515bdbbf..028516363 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/active_mask.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/active_mask.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operation for the generated active warp mask. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/clc.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/clc.rs index 4d251c358..f7f9f0bd6 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/clc.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/clc.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated Cluster Launch Control operations. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cluster_barrier.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cluster_barrier.rs index 73b38afd4..47116163a 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cluster_barrier.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cluster_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated operations for the closed cluster-barrier family. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cluster_memory.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cluster_memory.rs index b6a6c9984..83490c3b2 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cluster_memory.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cluster_memory.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for cluster address mapping and remote shared reads. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cp_async.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cp_async.rs index 553c161a5..6b381fdcc 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cp_async.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/cp_async.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for classic `cp.async` instructions. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/debug_control.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/debug_control.rs index 101fc85f8..5842b7f17 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/debug_control.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/debug_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for generated PTX debug controls. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/dotprod.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/dotprod.rs index 4acdd93b3..e9be20852 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/dotprod.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/dotprod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for generated packed integer dot products. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/elect.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/elect.rs index a8612264f..80e4e0b72 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/elect.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/elect.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated warp leader-election operation. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/execution_control.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/execution_control.rs index f998bbb13..e7f4eb825 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/execution_control.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/execution_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated counted-barrier, grid-dependency, and register-control operations. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/extended_minmax.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/extended_minmax.rs index e9f503bb0..83d99129d 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/extended_minmax.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/extended_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! One structural operation for extended floating-point min/max. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/integer_minmax.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/integer_minmax.rs index 7013c998e..963fb56a7 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/integer_minmax.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/integer_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for generated extended integer min/max. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/ldmatrix.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/ldmatrix.rs index 74ec271c1..4e5bfb1ee 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/ldmatrix.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/ldmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operation and compatibility carriers for generated `ldmatrix`. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mbarrier_basic.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mbarrier_basic.rs index ea844c8cf..59068a3db 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mbarrier_basic.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mbarrier_basic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for the basic shared-memory mbarrier lifecycle. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mbarrier_extended.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mbarrier_extended.rs index d5de90d50..fdf4c8fa5 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mbarrier_extended.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mbarrier_extended.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for extended mbarrier and async-proxy instructions. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mod.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mod.rs index 2cb7aff92..555ffa125 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mod.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/mod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. mod active_mask; diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/movmatrix.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/movmatrix.rs index 88980724c..0e8a6e9ce 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/movmatrix.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/movmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operation for generated in-register matrix transpose. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_alu.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_alu.rs index 51e660245..380726ac1 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_alu.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_alu.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for generated packed floating-point arithmetic. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_atomic.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_atomic.rs index 8c19c30ae..d608bad23 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_atomic.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_atomic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! One structural operation for the closed generated packed-atomic family. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_conversion.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_conversion.rs index fa83a5345..03eb96c07 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_conversion.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/packed_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operation for generated packed conversion. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/prmt.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/prmt.rs index f147341d0..2cd3e3b93 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/prmt.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/prmt.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! One structural operation for the closed generated `prmt` family. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/redux.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/redux.rs index c6bcdbf3b..c3914109d 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/redux.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/redux.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural NVVM operations for the closed generated `redux.sync` family. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/register_mma.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/register_mma.rs index 4767e487c..7bb191879 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/register_mma.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/register_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! One structural operation for generated register-only `mma.sync` variants. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_arithmetic.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_arithmetic.rs index fa498fbe9..dc1d8c0d3 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_arithmetic.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_arithmetic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! One structural operation for generated scalar arithmetic. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_conversion.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_conversion.rs index d6a767a80..4495afa83 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_conversion.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! One structural operation for generated `f32` to TF32 conversions. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_math.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_math.rs index 0cdce1daa..763ac4cb9 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_math.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/scalar_math.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! One structural operation for generated scalar math. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sparse_mma.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sparse_mma.rs index 9aaa2f04c..45386eda8 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sparse_mma.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sparse_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! One structural operation for generated sparse MMA variants. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sreg.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sreg.rs index f63a3efac..a6a8000b1 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sreg.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural NVVM operations for generated special-register reads. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/stmatrix.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/stmatrix.rs index 35fad6b58..40e487b4c 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/stmatrix.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/stmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for the four generated `stmatrix` stores. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sync.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sync.rs index 212c36f82..f58708c17 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sync.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/sync.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operation for generated CTA synchronization. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/tcgen05.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/tcgen05.rs index 88b913dea..be79474f4 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/tcgen05.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/tcgen05.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated Tensor Core Generation 5 operations. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/tma.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/tma.rs index de5e56756..4c100e828 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/tma.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/tma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated Tensor Memory Accelerator operations. @@ -227,6 +227,21 @@ impl CpAsyncBulkTensorG2sTile1dOp { } } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +#[pliron_op( + name = "nvvm.cp_async_bulk_tensor_g2s_cta_tile_1d", + format, + verifier = "succ", + interfaces = [NOpdsInterface<5>, NResultsInterface<0>], +)] +pub struct CpAsyncBulkTensorG2sCtaTile1dOp; + +impl CpAsyncBulkTensorG2sCtaTile1dOp { + pub fn new(op: Ptr) -> Self { + Self { op } + } +} + /// Starts a TMA tile copy from shared to global memory. #[pliron_op( name = "nvvm.cp_async_bulk_tensor_s2g_tile_1d", @@ -257,6 +272,21 @@ impl CpAsyncBulkTensorG2sTile2dOp { } } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +#[pliron_op( + name = "nvvm.cp_async_bulk_tensor_g2s_cta_tile_2d", + format, + verifier = "succ", + interfaces = [NOpdsInterface<6>, NResultsInterface<0>], +)] +pub struct CpAsyncBulkTensorG2sCtaTile2dOp; + +impl CpAsyncBulkTensorG2sCtaTile2dOp { + pub fn new(op: Ptr) -> Self { + Self { op } + } +} + /// Starts a multicast TMA tile copy from global to cluster shared memory. #[pliron_op( name = "nvvm.cp_async_bulk_tensor_g2s_tile_2d_multicast", @@ -317,6 +347,21 @@ impl CpAsyncBulkTensorG2sTile3dOp { } } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +#[pliron_op( + name = "nvvm.cp_async_bulk_tensor_g2s_cta_tile_3d", + format, + verifier = "succ", + interfaces = [NOpdsInterface<7>, NResultsInterface<0>], +)] +pub struct CpAsyncBulkTensorG2sCtaTile3dOp; + +impl CpAsyncBulkTensorG2sCtaTile3dOp { + pub fn new(op: Ptr) -> Self { + Self { op } + } +} + /// Starts a TMA tile copy from shared to global memory. #[pliron_op( name = "nvvm.cp_async_bulk_tensor_s2g_tile_3d", @@ -347,6 +392,21 @@ impl CpAsyncBulkTensorG2sTile4dOp { } } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +#[pliron_op( + name = "nvvm.cp_async_bulk_tensor_g2s_cta_tile_4d", + format, + verifier = "succ", + interfaces = [NOpdsInterface<8>, NResultsInterface<0>], +)] +pub struct CpAsyncBulkTensorG2sCtaTile4dOp; + +impl CpAsyncBulkTensorG2sCtaTile4dOp { + pub fn new(op: Ptr) -> Self { + Self { op } + } +} + /// Starts a TMA tile copy from shared to global memory. #[pliron_op( name = "nvvm.cp_async_bulk_tensor_s2g_tile_4d", @@ -377,6 +437,21 @@ impl CpAsyncBulkTensorG2sTile5dOp { } } +/// Starts a TMA tile copy from global to the issuing CTA's shared memory. +#[pliron_op( + name = "nvvm.cp_async_bulk_tensor_g2s_cta_tile_5d", + format, + verifier = "succ", + interfaces = [NOpdsInterface<9>, NResultsInterface<0>], +)] +pub struct CpAsyncBulkTensorG2sCtaTile5dOp; + +impl CpAsyncBulkTensorG2sCtaTile5dOp { + pub fn new(op: Ptr) -> Self { + Self { op } + } +} + /// Starts a TMA tile copy from shared to global memory. #[pliron_op( name = "nvvm.cp_async_bulk_tensor_s2g_tile_5d", @@ -1697,16 +1772,21 @@ pub(super) fn register(ctx: &mut Context) { CpAsyncBulkPrefetchTensorGather4TwoDimensionalL2Op::register(ctx); CpAsyncBulkPrefetchTensorGather4TwoDimensionalL2CacheHintOp::register(ctx); CpAsyncBulkTensorG2sTile1dOp::register(ctx); + CpAsyncBulkTensorG2sCtaTile1dOp::register(ctx); CpAsyncBulkTensorS2gTile1dOp::register(ctx); CpAsyncBulkTensorG2sTile2dOp::register(ctx); + CpAsyncBulkTensorG2sCtaTile2dOp::register(ctx); CpAsyncBulkTensorG2sTile2dMulticastOp::register(ctx); CpAsyncBulkTensorG2sTile2dMulticastCg2Op::register(ctx); CpAsyncBulkTensorS2gTile2dOp::register(ctx); CpAsyncBulkTensorG2sTile3dOp::register(ctx); + CpAsyncBulkTensorG2sCtaTile3dOp::register(ctx); CpAsyncBulkTensorS2gTile3dOp::register(ctx); CpAsyncBulkTensorG2sTile4dOp::register(ctx); + CpAsyncBulkTensorG2sCtaTile4dOp::register(ctx); CpAsyncBulkTensorS2gTile4dOp::register(ctx); CpAsyncBulkTensorG2sTile5dOp::register(ctx); + CpAsyncBulkTensorG2sCtaTile5dOp::register(ctx); CpAsyncBulkTensorS2gTile5dOp::register(ctx); CpAsyncBulkTensorReduceAddIm2col3dOp::register(ctx); CpAsyncBulkTensorReduceAddIm2col4dOp::register(ctx); diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/vote.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/vote.rs index 02e65111a..a14d51fe2 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/vote.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/vote.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for the generated `vote.sync` family. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_barrier.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_barrier.rs index 0698e06b2..196ebb1ed 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_barrier.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operation for generated warp synchronization. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_match.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_match.rs index 4472cf1ca..8593a8d9d 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_match.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_match.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for the generated `match.sync` family. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_shuffle.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_shuffle.rs index 611cf8e79..5b7121aae 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_shuffle.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/warp_shuffle.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for the generated `shfl.sync` family. diff --git a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/wgmma_control.rs b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/wgmma_control.rs index 5733dd83c..d18eb1b15 100644 --- a/cuda-oxide/crates/dialect-nvvm/src/ops/generated/wgmma_control.rs +++ b/cuda-oxide/crates/dialect-nvvm/src/ops/generated/wgmma_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Structural operations for generated WGMMA controls. diff --git a/cuda-oxide/crates/llvm-export/src/export/function.rs b/cuda-oxide/crates/llvm-export/src/export/function.rs index f8e6e1d4d..db8900f9a 100644 --- a/cuda-oxide/crates/llvm-export/src/export/function.rs +++ b/cuda-oxide/crates/llvm-export/src/export/function.rs @@ -65,6 +65,43 @@ fn sanitize_comment(text: &str) -> String { /// Public export callers need not have verified the source module. Pliron 0.19 /// also sums segment sizes as u32 without checking overflow, so validate the /// complete layout here before it can index or allocate from malformed counts. +/// Looks for a pointer into shared (address space 3) or local (address space +/// 5) memory inside a kernel parameter type, through struct fields and array +/// elements. On a hit it returns the address space and appends the path that +/// reaches the pointer to `path`: `.N` for struct field `N` and `[0]` for an +/// array element (every element has the same type, so the first stands for +/// all of them). A bare pointer leaves `path` empty. +fn device_only_pointer_in(ctx: &Context, ty: TypeHandle, path: &mut String) -> Option { + let ty_ref = ty.deref(ctx); + if let Some(pointer) = ty_ref.downcast_ref::() { + let space = pointer.address_space(); + return matches!(space, 3 | 5).then_some(space); + } + if let Some(array) = ty_ref.downcast_ref::() { + let mark = path.len(); + path.push_str("[0]"); + if let Some(space) = device_only_pointer_in(ctx, array.elem_type(), path) { + return Some(space); + } + path.truncate(mark); + return None; + } + if let Some(structure) = ty_ref.downcast_ref::() { + if structure.is_opaque() { + return None; + } + for (field_index, field) in structure.fields().enumerate() { + let mark = path.len(); + path.push_str(&format!(".{field_index}")); + if let Some(space) = device_only_pointer_in(ctx, field, path) { + return Some(space); + } + path.truncate(mark); + } + } + None +} + fn validate_conditional_branch_layout(ctx: &Context, branch: &ops::CondBrOp) -> Result<(), String> { let operation = branch.get_operation().deref(ctx); let sizes = operation @@ -999,23 +1036,40 @@ impl<'a> ModuleExportState<'a> { // parameter type today; if a frontend type ever surfaces it at a // kernel boundary, it needs a new arm here for the same reason as // plain shared. + // + // The check looks through struct fields and array elements as well. + // A struct or array parameter reaches the kernel as the bytes the + // host wrote, so a shared or local pointer inside one is as + // impossible to supply as a bare one. Before this recursion a + // `struct { p: SharedPtr, n: u32 }` parameter compiled and the + // kernel applied `cvta.shared.u64` to host bytes. if is_kernel { for (index, arg_ty) in func_ty.arg_types().iter().enumerate() { - let arg_ref = arg_ty.deref(self.ctx); - let Some(pointer) = arg_ref.downcast_ref::() else { + let mut path = String::new(); + let Some(address_space) = device_only_pointer_in(self.ctx, *arg_ty, &mut path) + else { continue; }; - let space = match pointer.address_space() { - 3 => "shared", - 5 => "local", - _ => continue, + let space = if address_space == 3 { + "shared" + } else { + "local" }; + let reason = format!( + "{space} memory is allocated by the device at launch, so the host has no \ + address to pass, and the driver refuses a module whose entry declares one" + ); + if path.is_empty() { + return Err(format!( + "kernel `@{fixed_func_name}` parameter {index} is a pointer into {space} \ + memory (address space {address_space}); {reason}" + )); + } + // The field path keeps this message apart from the bare form, + // so a fixture can pin the nested case on its own. return Err(format!( - "kernel `@{fixed_func_name}` parameter {index} is a pointer into {space} \ - memory (address space {}); {space} memory is allocated by the device at \ - launch, so the host has no address to pass, and the driver refuses a \ - module whose entry declares one", - pointer.address_space() + "kernel `@{fixed_func_name}` parameter {index} holds a pointer into {space} \ + memory (address space {address_space}) at field path `{path}`; {reason}" )); } } diff --git a/cuda-oxide/crates/llvm-export/tests/export_test/kernels_and_roots.rs b/cuda-oxide/crates/llvm-export/tests/export_test/kernels_and_roots.rs index c557166df..d80832d0a 100644 --- a/cuda-oxide/crates/llvm-export/tests/export_test/kernels_and_roots.rs +++ b/cuda-oxide/crates/llvm-export/tests/export_test/kernels_and_roots.rs @@ -10,7 +10,7 @@ use llvm_export::{ export_module_with_externs_and_roots, }, ops::{FuncOp, GlobalOp, GlobalOpExt, ReturnOp}, - types::{FuncType, PointerType, VoidType}, + types::{ArrayType, FuncType, PointerType, StructLayout, StructType, VoidType}, }; use pliron::{ basic_block::BasicBlock, @@ -22,6 +22,7 @@ use pliron::{ context::{Context, Ptr}, identifier::Identifier, op::Op, + r#type::TypeHandle, }; use crate::common::module_top_block; @@ -246,10 +247,22 @@ fn pointer_param_function( address_spaces: &[u32], is_kernel: bool, ) { - let params: Vec<_> = address_spaces + let params: Vec = address_spaces .iter() .map(|space| PointerType::get(ctx, *space).into()) .collect(); + typed_param_function(ctx, module_block, name, params, is_kernel); +} + +/// Builds a `void` function with the given parameter types and an empty +/// body, optionally carrying the `gpu_kernel` marker. +fn typed_param_function( + ctx: &mut Context, + module_block: Ptr, + name: &str, + params: Vec, + is_kernel: bool, +) { let void_ty = VoidType::get(ctx); let func_ty = FuncType::get(ctx, void_ty.into(), params, false); let func = FuncOp::new(ctx, name.try_into().unwrap(), func_ty); @@ -307,6 +320,165 @@ fn kernel_rejects_a_local_memory_pointer_parameter() { ); } +/// The host fills a struct parameter byte by byte, so a shared pointer +/// inside one is as impossible to supply as a bare one. The exporter looks +/// through the aggregate and names the field path it found, with a message +/// distinct from the bare form. +#[test] +fn kernel_rejects_a_shared_memory_pointer_inside_a_struct_parameter() { + let mut ctx = Context::new(); + let module = ModuleOp::new(&mut ctx, "kernel_shared_struct_param".try_into().unwrap()); + let module_block = module_top_block(&mut ctx, &module); + let shared_ptr_ty = PointerType::get(&ctx, 3); + let i32_ty = IntegerType::get(&ctx, 32, Signedness::Signless); + let args_ty = StructType::get_unnamed( + &ctx, + ( + vec![shared_ptr_ty.into(), i32_ty.into()], + StructLayout::Unpacked, + ), + ); + typed_param_function( + &mut ctx, + module_block, + "shared_struct_param", + vec![args_ty.into()], + true, + ); + + let error = export_module_to_string(&ctx, &module) + .expect_err("export must refuse a kernel parameter struct that holds a shared pointer"); + assert!( + error.contains( + "kernel `@shared_struct_param` parameter 0 holds a pointer into shared memory \ + (address space 3) at field path `.0`" + ), + "{error}" + ); +} + +/// An array parameter is refused the same way, and the path names the +/// element position. +#[test] +fn kernel_rejects_a_shared_memory_pointer_inside_an_array_parameter() { + let mut ctx = Context::new(); + let module = ModuleOp::new(&mut ctx, "kernel_shared_array_param".try_into().unwrap()); + let module_block = module_top_block(&mut ctx, &module); + let generic_ptr_ty = PointerType::get(&ctx, 0); + let shared_ptr_ty = PointerType::get(&ctx, 3); + let pair_ty = ArrayType::get(&ctx, shared_ptr_ty.into(), 2); + typed_param_function( + &mut ctx, + module_block, + "shared_array_param", + vec![generic_ptr_ty.into(), pair_ty.into()], + true, + ); + + let error = export_module_to_string(&ctx, &module) + .expect_err("export must refuse a kernel parameter array of shared pointers"); + assert!( + error.contains( + "kernel `@shared_array_param` parameter 1 holds a pointer into shared memory \ + (address space 3) at field path `[0]`" + ), + "{error}" + ); +} + +/// A pointer buried under a struct, an array and another struct is found and +/// reported with its full path; local memory is refused on the same grounds. +#[test] +fn kernel_rejects_a_local_memory_pointer_nested_below_an_array_element() { + let mut ctx = Context::new(); + let module = ModuleOp::new(&mut ctx, "kernel_local_nested_param".try_into().unwrap()); + let module_block = module_top_block(&mut ctx, &module); + let local_ptr_ty = PointerType::get(&ctx, 5); + let i32_ty = IntegerType::get(&ctx, 32, Signedness::Signless); + let leaf_ty = StructType::get_unnamed( + &ctx, + ( + vec![local_ptr_ty.into(), i32_ty.into()], + StructLayout::Unpacked, + ), + ); + let pair_ty = ArrayType::get(&ctx, leaf_ty.into(), 2); + let args_ty = StructType::get_unnamed( + &ctx, + (vec![i32_ty.into(), pair_ty.into()], StructLayout::Unpacked), + ); + typed_param_function( + &mut ctx, + module_block, + "local_nested_param", + vec![args_ty.into()], + true, + ); + + let error = export_module_to_string(&ctx, &module) + .expect_err("export must refuse a nested local-memory pointer in a kernel parameter"); + assert!( + error.contains( + "kernel `@local_nested_param` parameter 0 holds a pointer into local memory \ + (address space 5) at field path `.1[0].0`" + ), + "{error}" + ); +} + +/// The recursion refuses only what the host cannot supply: a struct of +/// generic and global pointers stays a valid kernel parameter, and a device +/// (non-kernel) function may take the struct that holds a shared pointer. +#[test] +fn kernel_keeps_a_struct_of_host_addressable_pointers_and_device_functions_keep_nested_shared() { + let mut ctx = Context::new(); + let module = ModuleOp::new(&mut ctx, "kernel_allowed_struct_params".try_into().unwrap()); + let module_block = module_top_block(&mut ctx, &module); + let generic_ptr_ty = PointerType::get(&ctx, 0); + let global_ptr_ty = PointerType::get(&ctx, 1); + let shared_ptr_ty = PointerType::get(&ctx, 3); + let i32_ty = IntegerType::get(&ctx, 32, Signedness::Signless); + let host_ty = StructType::get_unnamed( + &ctx, + ( + vec![generic_ptr_ty.into(), global_ptr_ty.into()], + StructLayout::Unpacked, + ), + ); + let shared_ty = StructType::get_unnamed( + &ctx, + ( + vec![shared_ptr_ty.into(), i32_ty.into()], + StructLayout::Unpacked, + ), + ); + typed_param_function( + &mut ctx, + module_block, + "host_struct_param", + vec![host_ty.into()], + true, + ); + typed_param_function( + &mut ctx, + module_block, + "device_shared_struct", + vec![shared_ty.into()], + false, + ); + + let ir = export_module_to_string(&ctx, &module) + .expect("a host-addressable struct parameter and a device function's shared struct export"); + assert!( + ir.contains("define ptx_kernel void @host_struct_param({ ptr, ptr addrspace(1) } %v0)"), + "{ir}" + ); + assert!( + ir.contains("@device_shared_struct({ ptr addrspace(3), i32 } %v0)"), + "{ir}" + ); +} + /// The refusal is scoped to what the host cannot supply. Generic, global, /// and constant pointer parameters are host-addressable and stay accepted on /// kernels, and a device (non-kernel) function may carry any state space on diff --git a/cuda-oxide/crates/mir-importer/src/translator/facts.rs b/cuda-oxide/crates/mir-importer/src/translator/facts.rs index db134cb78..c9542c38f 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/facts.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/facts.rs @@ -417,6 +417,15 @@ mod pointer_origin { } } + /// ABI RULE: `cuda_device::SharedPtr` stores a `*mut SharedSlot`, so + /// the shared pointer it lowers to is always `RawMut`. + pub(crate) fn abi_shared_ptr_field() -> PointerOrigin { + PointerOrigin { + kind: MirPointerKind::RawMut, + mutable: true, + } + } + /// ABI RULE: `cuda_device::DisjointSlice<'_, T>` stores its data pointer /// as `*mut T`, so its data/element pointers are always `RawMut`. pub(crate) fn abi_disjoint_slice_data_ptr() -> PointerOrigin { diff --git a/cuda-oxide/crates/mir-importer/src/translator/rvalue/place_addr.rs b/cuda-oxide/crates/mir-importer/src/translator/rvalue/place_addr.rs index 21cecc504..b5c109733 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/rvalue/place_addr.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/rvalue/place_addr.rs @@ -64,6 +64,7 @@ pub(crate) fn translate_place_address( prev_op: Option>, loc: Location, ) -> TranslationResult>)>> { + let place = &*super::place_read::strip_shared_ptr_field_projections(body, place); let Some(slot) = value_map.get_slot(place.local) else { return Ok(None); }; diff --git a/cuda-oxide/crates/mir-importer/src/translator/rvalue/place_read.rs b/cuda-oxide/crates/mir-importer/src/translator/rvalue/place_read.rs index 3ec3693a5..dc3f96e9d 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/rvalue/place_read.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/rvalue/place_read.rs @@ -80,6 +80,7 @@ pub fn translate_place( prev_op: Option>, loc: Location, ) -> TranslationResult<(Value, Option>)> { + let place = &*strip_shared_ptr_field_projections(body, place); match classify_place_read_strategy(ctx, place, value_map)? { PlaceReadStrategy::Address => { if let Some((value, last_op)) = translate_place_load_from_address( @@ -109,6 +110,59 @@ pub fn translate_place( } } +/// Drop the projections of a place onto the field of a `cuda_device::SharedPtr`. +/// +/// `SharedPtr { ptr: *mut SharedSlot }` is a scalar-lowered newtype: its +/// value is its one field, a CTA-shared pointer (see `types::translate_type`). +/// Projecting that field is therefore the identity. Without this, the place +/// walkers would read the pointer value of a `SharedPtr` nested in another +/// place, such as `(region.base).ptr`, as the address of an aggregate. +/// +/// `ThreadIndex`, the other scalar-lowered newtype, needs no strip. Its scalar +/// is an integer, and `dialect-mir` and `mir-lower` already treat +/// `extract_field` on an integer scalar as the identity. The iterative place +/// walker in `place_iter.rs` takes any pointer-typed current value as an +/// address to load from, so only a newtype whose scalar is itself a pointer is +/// misread. Follow-up: have that walker track whether its current value is an +/// address or a value, then drop this strip. +pub(crate) fn strip_shared_ptr_field_projections<'a>( + body: &mir::Body, + place: &'a mir::Place, +) -> std::borrow::Cow<'a, mir::Place> { + use rustc_public::ty::{RigidTy, TyKind}; + + if !place + .projection + .iter() + .any(|elem| matches!(elem, mir::ProjectionElem::Field(0, _))) + { + return std::borrow::Cow::Borrowed(place); + } + let mut place_ty = body.locals()[place.local].ty; + let mut kept = Vec::with_capacity(place.projection.len()); + for elem in &place.projection { + let is_shared_ptr = matches!( + place_ty.kind(), + TyKind::RigidTy(RigidTy::Adt(adt_def, _)) if facts::is_cuda_device_adt(&adt_def, "SharedPtr") + ); + let Ok(next_ty) = elem.ty(place_ty) else { + return std::borrow::Cow::Borrowed(place); + }; + if !(is_shared_ptr && matches!(elem, mir::ProjectionElem::Field(0, _))) { + kept.push(elem.clone()); + } + place_ty = next_ty; + } + if kept.len() == place.projection.len() { + std::borrow::Cow::Borrowed(place) + } else { + std::borrow::Cow::Owned(mir::Place { + local: place.local, + projection: kept, + }) + } +} + // ============================================================================ // Place Read Strategy and Address Path // ============================================================================ diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/active_mask.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/active_mask.rs index ce0a280e2..07bd68d3e 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/active_mask.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/active_mask.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `active_mask` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/clc.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/clc.rs index a1e5adb5a..c8e4d5bae 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/clc.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/clc.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `clc` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cluster_barrier.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cluster_barrier.rs index 05c3918a4..8e0078ab1 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cluster_barrier.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cluster_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `cluster_barrier` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cluster_memory.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cluster_memory.rs index c11ddf610..28539f3b1 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cluster_memory.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cluster_memory.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `cluster_memory` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cp_async.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cp_async.rs index 3c840f9b5..191abcfbc 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cp_async.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/cp_async.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `cp_async` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/debug_control.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/debug_control.rs index a9e4e19ff..e132908a9 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/debug_control.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/debug_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `debug_control` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/dotprod.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/dotprod.rs index 3e7c7620d..69ef08528 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/dotprod.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/dotprod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `dotprod` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/elect.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/elect.rs index 59f961cb3..e6689bcb4 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/elect.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/elect.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `elect` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/execution_control.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/execution_control.rs index ee0c6b755..5d882c857 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/execution_control.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/execution_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `execution_control` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/extended_minmax.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/extended_minmax.rs index 967a90188..fbc2004a4 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/extended_minmax.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/extended_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `extended_minmax` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/helpers.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/helpers.rs index 9a1899a62..1f0b35910 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/helpers.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/helpers.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Shared glue for the generated intrinsic dispatch shards. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/integer_minmax.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/integer_minmax.rs index 6f3fdd86f..2f5d55c53 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/integer_minmax.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/integer_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `integer_minmax` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/ldmatrix.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/ldmatrix.rs index 7dfe1ccfb..32efb100e 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/ldmatrix.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/ldmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `ldmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mbarrier_basic.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mbarrier_basic.rs index eb53af9b1..20da6ede9 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mbarrier_basic.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mbarrier_basic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `mbarrier_basic` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mbarrier_extended.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mbarrier_extended.rs index ad552194d..cc0182a01 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mbarrier_extended.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mbarrier_extended.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `mbarrier_extended` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mod.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mod.rs index 9c980fca9..6b65a34d5 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mod.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/mod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility path dispatch for CUDA intrinsics. @@ -313,10 +313,14 @@ pub fn is_generated_intrinsic_path(name: &str) -> bool { | "cuda_device::tma::cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0328" | "cuda_device::tma::cp_async_bulk_tensor_1d_g2s" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034" + | "cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0335" | "cuda_device::tma::cp_async_bulk_tensor_1d_s2g" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0329" | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035" + | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0330" | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_multicast" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0331" @@ -325,14 +329,20 @@ pub fn is_generated_intrinsic_path(name: &str) -> bool { | "cuda_device::tma::cp_async_bulk_tensor_2d_s2g" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0332" | "cuda_device::tma::cp_async_bulk_tensor_3d_g2s" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036" + | "cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337" | "cuda_device::tma::cp_async_bulk_tensor_3d_s2g" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0333" | "cuda_device::tma::cp_async_bulk_tensor_4d_g2s" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037" + | "cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338" | "cuda_device::tma::cp_async_bulk_tensor_4d_s2g" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0334" | "cuda_device::tma::cp_async_bulk_tensor_5d_g2s" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038" + | "cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0339" | "cuda_device::tma::cp_async_bulk_tensor_5d_s2g" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0928" @@ -2309,16 +2319,21 @@ pub fn is_raw_generated_intrinsic_path(name: &str) -> bool { | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0893" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0922" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0328" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0335" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0329" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0330" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0331" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0336" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0332" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0333" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0334" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0339" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0928" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0929" @@ -3443,10 +3458,14 @@ pub fn generated_intrinsic_marker(name: &str) -> Option<&'static str> { } "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0328" | "cuda_device::tma::cp_async_bulk_tensor_1d_g2s" => Some("v1:i0328"), + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034" + | "cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta" => Some("v1:i1034"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0335" | "cuda_device::tma::cp_async_bulk_tensor_1d_s2g" => Some("v1:i0335"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0329" | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s" => Some("v1:i0329"), + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035" + | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta" => Some("v1:i1035"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0330" | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_multicast" => Some("v1:i0330"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0331" @@ -3455,14 +3474,20 @@ pub fn generated_intrinsic_marker(name: &str) -> Option<&'static str> { | "cuda_device::tma::cp_async_bulk_tensor_2d_s2g" => Some("v1:i0336"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0332" | "cuda_device::tma::cp_async_bulk_tensor_3d_g2s" => Some("v1:i0332"), + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036" + | "cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta" => Some("v1:i1036"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337" | "cuda_device::tma::cp_async_bulk_tensor_3d_s2g" => Some("v1:i0337"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0333" | "cuda_device::tma::cp_async_bulk_tensor_4d_g2s" => Some("v1:i0333"), + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037" + | "cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta" => Some("v1:i1037"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338" | "cuda_device::tma::cp_async_bulk_tensor_4d_s2g" => Some("v1:i0338"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0334" | "cuda_device::tma::cp_async_bulk_tensor_5d_g2s" => Some("v1:i0334"), + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038" + | "cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta" => Some("v1:i1038"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0339" | "cuda_device::tma::cp_async_bulk_tensor_5d_s2g" => Some("v1:i0339"), "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0928" diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/movmatrix.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/movmatrix.rs index 2384dc6f8..58fbc4a63 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/movmatrix.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/movmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `movmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_alu.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_alu.rs index 480526765..327b3af7c 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_alu.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_alu.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `packed_alu` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_atomic.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_atomic.rs index 32377c7d4..b17127860 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_atomic.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_atomic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `packed_atomic` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_conversion.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_conversion.rs index b7cbacf93..76f72e521 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_conversion.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/packed_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `packed_conversion` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/prmt.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/prmt.rs index a15e3196a..b1bb0fc01 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/prmt.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/prmt.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `prmt` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/redux.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/redux.rs index 9a630b3ed..918c06573 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/redux.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/redux.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `redux` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/register_mma.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/register_mma.rs index 47ad59a70..084dc6b73 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/register_mma.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/register_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `register_mma` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_arithmetic.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_arithmetic.rs index cbdf2bbda..addefa3aa 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_arithmetic.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_arithmetic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `scalar_arithmetic` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_conversion.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_conversion.rs index 9e08cbdc7..2bb6ea0f9 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_conversion.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `scalar_conversion` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_math.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_math.rs index 93fc0784e..faac1b40c 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_math.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/scalar_math.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `scalar_math` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sparse_mma.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sparse_mma.rs index 993ac1a84..aeb6f4692 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sparse_mma.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sparse_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `sparse_mma` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sreg.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sreg.rs index 97c70d888..ecb4a0002 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sreg.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `sreg` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/stmatrix.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/stmatrix.rs index a0f41e97f..0fab7612c 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/stmatrix.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/stmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `stmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sync.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sync.rs index 484c1a6a7..883e6627d 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sync.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/sync.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `sync` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_cp.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_cp.rs index 9aa3a1c92..859850ed4 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_cp.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_cp.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `tcgen05_cp` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_ld.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_ld.rs index d47e6abb1..ea74b7936 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_ld.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_ld.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `tcgen05_ld` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_mma.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_mma.rs index eec859a00..73b27e24b 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_mma.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `tcgen05_mma` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_other.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_other.rs index 223a7e211..717c58f43 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_other.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_other.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `tcgen05_other` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_st.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_st.rs index 599f8c5df..3feab3a35 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_st.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tcgen05_st.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `tcgen05_st` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/active_mask.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/active_mask.rs index c24ccc905..e043060f7 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/active_mask.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/active_mask.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `active_mask` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/clc.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/clc.rs index da599ce05..23dd6691e 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/clc.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/clc.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `clc` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cluster_barrier.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cluster_barrier.rs index a96747bd8..722175dad 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cluster_barrier.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cluster_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `cluster_barrier` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cluster_memory.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cluster_memory.rs index 6218b18f6..04ef58485 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cluster_memory.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cluster_memory.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `cluster_memory` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cp_async.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cp_async.rs index 5b1499015..4bba62c61 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cp_async.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/cp_async.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `cp_async` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/debug_control.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/debug_control.rs index 160cdd254..2fed06b33 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/debug_control.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/debug_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `debug_control` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/dotprod.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/dotprod.rs index 481dcfbe0..0c0e6867f 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/dotprod.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/dotprod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `dotprod` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/elect.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/elect.rs index c73be2a52..d896c9f2c 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/elect.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/elect.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `elect` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/execution_control.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/execution_control.rs index 57351d2b9..3195737ca 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/execution_control.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/execution_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `execution_control` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/extended_minmax.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/extended_minmax.rs index b7ce30d14..375e65af3 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/extended_minmax.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/extended_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `extended_minmax` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/integer_minmax.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/integer_minmax.rs index 87f98b4df..65de75d12 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/integer_minmax.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/integer_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `integer_minmax` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/ldmatrix.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/ldmatrix.rs index da129938f..1631dfdc2 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/ldmatrix.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/ldmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `ldmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mbarrier_basic.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mbarrier_basic.rs index a6235ba46..462d98f01 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mbarrier_basic.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mbarrier_basic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `mbarrier_basic` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mbarrier_extended.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mbarrier_extended.rs index 45f494fe2..9a85710cc 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mbarrier_extended.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mbarrier_extended.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `mbarrier_extended` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mod.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mod.rs index 65795232e..d6c89b19c 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mod.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/mod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests, grouped per intrinsic family. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/movmatrix.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/movmatrix.rs index cb5554095..df2e3cc86 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/movmatrix.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/movmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `movmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_alu.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_alu.rs index 1f01e82de..c3aabbec7 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_alu.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_alu.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `packed_alu` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_atomic.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_atomic.rs index 05b199090..6966a2e99 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_atomic.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_atomic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `packed_atomic` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_conversion.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_conversion.rs index 311c4f970..55f8fe447 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_conversion.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/packed_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `packed_conversion` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/prmt.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/prmt.rs index 0928fdfbc..c84bb6190 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/prmt.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/prmt.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `prmt` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/redux.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/redux.rs index cbe3751ad..0d242fcab 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/redux.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/redux.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `redux` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/register_mma.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/register_mma.rs index 38a894489..ed8e0f695 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/register_mma.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/register_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `register_mma` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_arithmetic.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_arithmetic.rs index 11eebceed..11f5adb32 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_arithmetic.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_arithmetic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `scalar_arithmetic` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_conversion.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_conversion.rs index 35a1416a2..348576302 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_conversion.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `scalar_conversion` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_math.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_math.rs index a6efe97fe..2258c54f8 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_math.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/scalar_math.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `scalar_math` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sparse_mma.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sparse_mma.rs index 375f4cddd..360724777 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sparse_mma.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sparse_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `sparse_mma` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sreg.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sreg.rs index 62cb47f77..d9f6e3e46 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sreg.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `sreg` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/stmatrix.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/stmatrix.rs index 7d2caa1ac..484742f42 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/stmatrix.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/stmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `stmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sync.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sync.rs index 0dad5bf19..1ea02524f 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sync.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/sync.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `sync` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/tcgen05.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/tcgen05.rs index 3eb2414b7..71ab9a684 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/tcgen05.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/tcgen05.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `tcgen05` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/tma.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/tma.rs index 7cac99888..f9fea098f 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/tma.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/tma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `tma` intrinsics. @@ -529,6 +529,42 @@ fn cp_async_bulk_tensor_1d_g2s_uses_only_canonical_or_compatibility_defpaths() { )); } #[test] +fn cp_async_bulk_tensor_1d_g2s_cta_uses_only_canonical_or_compatibility_defpaths() { + assert!(is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034" + )); + assert!(is_raw_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034" + )); + assert_eq!( + generated_intrinsic_marker("cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034"), + Some("v1:i1034") + ); + assert!(matches!( + classify_raw_intrinsic_path( + "cuda_intrinsics", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034".into() + ), + RawIntrinsicIdentity::Known(_) + )); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::tma::cp_async_bulk_tensor_1d_g2s_cta" + )); + assert!(is_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta" + )); + assert!(!is_raw_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta" + )); + assert_eq!( + generated_intrinsic_marker("cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta"), + Some("v1:i1034") + ); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v2::i1034" + )); +} +#[test] fn cp_async_bulk_tensor_1d_s2g_uses_only_canonical_or_compatibility_defpaths() { assert!(is_generated_intrinsic_path( "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0335" @@ -601,6 +637,42 @@ fn cp_async_bulk_tensor_2d_g2s_uses_only_canonical_or_compatibility_defpaths() { )); } #[test] +fn cp_async_bulk_tensor_2d_g2s_cta_uses_only_canonical_or_compatibility_defpaths() { + assert!(is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035" + )); + assert!(is_raw_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035" + )); + assert_eq!( + generated_intrinsic_marker("cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035"), + Some("v1:i1035") + ); + assert!(matches!( + classify_raw_intrinsic_path( + "cuda_intrinsics", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035".into() + ), + RawIntrinsicIdentity::Known(_) + )); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_cta" + )); + assert!(is_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta" + )); + assert!(!is_raw_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta" + )); + assert_eq!( + generated_intrinsic_marker("cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta"), + Some("v1:i1035") + ); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v2::i1035" + )); +} +#[test] fn cp_async_bulk_tensor_2d_g2s_multicast_uses_only_canonical_or_compatibility_defpaths() { assert!(is_generated_intrinsic_path( "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0330" @@ -745,6 +817,42 @@ fn cp_async_bulk_tensor_3d_g2s_uses_only_canonical_or_compatibility_defpaths() { )); } #[test] +fn cp_async_bulk_tensor_3d_g2s_cta_uses_only_canonical_or_compatibility_defpaths() { + assert!(is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036" + )); + assert!(is_raw_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036" + )); + assert_eq!( + generated_intrinsic_marker("cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036"), + Some("v1:i1036") + ); + assert!(matches!( + classify_raw_intrinsic_path( + "cuda_intrinsics", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036".into() + ), + RawIntrinsicIdentity::Known(_) + )); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s_cta" + )); + assert!(is_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta" + )); + assert!(!is_raw_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta" + )); + assert_eq!( + generated_intrinsic_marker("cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta"), + Some("v1:i1036") + ); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v2::i1036" + )); +} +#[test] fn cp_async_bulk_tensor_3d_s2g_uses_only_canonical_or_compatibility_defpaths() { assert!(is_generated_intrinsic_path( "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337" @@ -817,6 +925,42 @@ fn cp_async_bulk_tensor_4d_g2s_uses_only_canonical_or_compatibility_defpaths() { )); } #[test] +fn cp_async_bulk_tensor_4d_g2s_cta_uses_only_canonical_or_compatibility_defpaths() { + assert!(is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037" + )); + assert!(is_raw_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037" + )); + assert_eq!( + generated_intrinsic_marker("cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037"), + Some("v1:i1037") + ); + assert!(matches!( + classify_raw_intrinsic_path( + "cuda_intrinsics", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037".into() + ), + RawIntrinsicIdentity::Known(_) + )); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s_cta" + )); + assert!(is_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta" + )); + assert!(!is_raw_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta" + )); + assert_eq!( + generated_intrinsic_marker("cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta"), + Some("v1:i1037") + ); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v2::i1037" + )); +} +#[test] fn cp_async_bulk_tensor_4d_s2g_uses_only_canonical_or_compatibility_defpaths() { assert!(is_generated_intrinsic_path( "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338" @@ -889,6 +1033,42 @@ fn cp_async_bulk_tensor_5d_g2s_uses_only_canonical_or_compatibility_defpaths() { )); } #[test] +fn cp_async_bulk_tensor_5d_g2s_cta_uses_only_canonical_or_compatibility_defpaths() { + assert!(is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038" + )); + assert!(is_raw_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038" + )); + assert_eq!( + generated_intrinsic_marker("cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038"), + Some("v1:i1038") + ); + assert!(matches!( + classify_raw_intrinsic_path( + "cuda_intrinsics", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038".into() + ), + RawIntrinsicIdentity::Known(_) + )); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s_cta" + )); + assert!(is_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta" + )); + assert!(!is_raw_generated_intrinsic_path( + "cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta" + )); + assert_eq!( + generated_intrinsic_marker("cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta"), + Some("v1:i1038") + ); + assert!(!is_generated_intrinsic_path( + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v2::i1038" + )); +} +#[test] fn cp_async_bulk_tensor_5d_s2g_uses_only_canonical_or_compatibility_defpaths() { assert!(is_generated_intrinsic_path( "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0339" diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/vote.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/vote.rs index 26513fe1f..12870d14f 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/vote.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/vote.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `vote` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_barrier.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_barrier.rs index 1f5103a1e..2c2b5aec2 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_barrier.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `warp_barrier` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_match.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_match.rs index 8db832c68..2ab93890d 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_match.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_match.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `warp_match` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_shuffle.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_shuffle.rs index f8f1fd992..038ad862e 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_shuffle.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/warp_shuffle.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `warp_shuffle` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/wgmma_control.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/wgmma_control.rs index 97d4b2cac..834732137 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/wgmma_control.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tests/wgmma_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated dispatch-path tests: `wgmma_control` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tma.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tma.rs index 8059ee688..579000521 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tma.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/tma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `tma` intrinsics. @@ -587,6 +587,13 @@ pub(super) fn try_dispatch( "v1:i0328", )?)) } + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034" + | "cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta" => { + Ok(Some(super::super::tma::emit_tma_g2s_cta( + ctx, body, args, target, block_ptr, prev_op, value_map, block_map, loc, 1, + "v1:i1034", + )?)) + } "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0335" | "cuda_device::tma::cp_async_bulk_tensor_1d_s2g" => { Ok(Some(super::super::tma::emit_tma_s2g( @@ -601,6 +608,13 @@ pub(super) fn try_dispatch( "v1:i0329", )?)) } + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035" + | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta" => { + Ok(Some(super::super::tma::emit_tma_g2s_cta( + ctx, body, args, target, block_ptr, prev_op, value_map, block_map, loc, 2, + "v1:i1035", + )?)) + } "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0330" | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_multicast" => { Ok(Some(super::super::tma::emit_tma_g2s_multicast( @@ -627,6 +641,13 @@ pub(super) fn try_dispatch( "v1:i0332", )?)) } + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036" + | "cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta" => { + Ok(Some(super::super::tma::emit_tma_g2s_cta( + ctx, body, args, target, block_ptr, prev_op, value_map, block_map, loc, 3, + "v1:i1036", + )?)) + } "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337" | "cuda_device::tma::cp_async_bulk_tensor_3d_s2g" => { Ok(Some(super::super::tma::emit_tma_s2g( @@ -641,6 +662,13 @@ pub(super) fn try_dispatch( "v1:i0333", )?)) } + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037" + | "cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta" => { + Ok(Some(super::super::tma::emit_tma_g2s_cta( + ctx, body, args, target, block_ptr, prev_op, value_map, block_map, loc, 4, + "v1:i1037", + )?)) + } "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338" | "cuda_device::tma::cp_async_bulk_tensor_4d_s2g" => { Ok(Some(super::super::tma::emit_tma_s2g( @@ -655,6 +683,13 @@ pub(super) fn try_dispatch( "v1:i0334", )?)) } + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038" + | "cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta" => { + Ok(Some(super::super::tma::emit_tma_g2s_cta( + ctx, body, args, target, block_ptr, prev_op, value_map, block_map, loc, 5, + "v1:i1038", + )?)) + } "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0339" | "cuda_device::tma::cp_async_bulk_tensor_5d_s2g" => { Ok(Some(super::super::tma::emit_tma_s2g( diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/vote.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/vote.rs index 5cbe5f53a..68e89bfbc 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/vote.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/vote.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `vote` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_barrier.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_barrier.rs index c4e3307b0..7da22f606 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_barrier.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `warp_barrier` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_match.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_match.rs index 8b562adb4..ad7495855 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_match.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_match.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `warp_match` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_shuffle.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_shuffle.rs index bbb15ace9..a148cfb70 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_shuffle.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/warp_shuffle.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `warp_shuffle` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/wgmma_control.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/wgmma_control.rs index a9d7fadeb..b0b8cf543 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/wgmma_control.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/generated/wgmma_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated raw/compatibility dispatch arms: `wgmma_control` intrinsics. diff --git a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/tma.rs b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/tma.rs index 3f111e32e..034567db7 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/tma.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/terminator/intrinsics/tma.rs @@ -13,7 +13,9 @@ use crate::translator::rvalue; use crate::translator::values::ValueMap; use dialect_mir::ops::MirConstantOp; use dialect_nvvm::ops::{ - CpAsyncBulkCommitGroupOp, CpAsyncBulkTensorG2sTile1dOp, + CpAsyncBulkCommitGroupOp, CpAsyncBulkTensorG2sCtaTile1dOp, CpAsyncBulkTensorG2sCtaTile2dOp, + CpAsyncBulkTensorG2sCtaTile3dOp, CpAsyncBulkTensorG2sCtaTile4dOp, + CpAsyncBulkTensorG2sCtaTile5dOp, CpAsyncBulkTensorG2sTile1dOp, CpAsyncBulkTensorG2sTile2dMulticastCg2Op, CpAsyncBulkTensorG2sTile2dMulticastOp, CpAsyncBulkTensorG2sTile2dOp, CpAsyncBulkTensorG2sTile3dOp, CpAsyncBulkTensorG2sTile4dOp, CpAsyncBulkTensorG2sTile5dOp, CpAsyncBulkTensorS2gTile1dOp, CpAsyncBulkTensorS2gTile2dOp, @@ -61,6 +63,44 @@ pub fn emit_tma_g2s( loc: Location, dims: usize, marker: &str, +) -> TranslationResult> { + emit_tma_g2s_impl( + ctx, body, args, target, block_ptr, prev_op, value_map, block_map, loc, dims, marker, false, + ) +} + +/// Emits a TMA G2S form whose destination belongs to the issuing CTA. +pub fn emit_tma_g2s_cta( + ctx: &mut Context, + body: &mir::Body, + args: &[mir::Operand], + target: &Option, + block_ptr: Ptr, + prev_op: Option>, + value_map: &mut ValueMap, + block_map: &[Ptr], + loc: Location, + dims: usize, + marker: &str, +) -> TranslationResult> { + emit_tma_g2s_impl( + ctx, body, args, target, block_ptr, prev_op, value_map, block_map, loc, dims, marker, true, + ) +} + +fn emit_tma_g2s_impl( + ctx: &mut Context, + body: &mir::Body, + args: &[mir::Operand], + target: &Option, + block_ptr: Ptr, + prev_op: Option>, + value_map: &mut ValueMap, + block_map: &[Ptr], + loc: Location, + dims: usize, + marker: &str, + cta: bool, ) -> TranslationResult> { // Expected args: dst, tensor_map, coord0, [coord1, ...], barrier let expected_args = 3 + dims; // dst + tensor_map + coords + barrier @@ -68,8 +108,9 @@ pub fn emit_tma_g2s( return input_err!( loc.clone(), TranslationErr::unsupported(format!( - "cp_async_bulk_tensor_{}d_g2s expects {} arguments, got {}", + "cp_async_bulk_tensor_{}d_g2s{} expects {} arguments, got {}", dims, + if cta { "_cta" } else { "" }, expected_args, args.len() )) @@ -135,38 +176,38 @@ pub fn emit_tma_g2s( last_op = last_op_after; } - // Add default cta_mask (i16 = 0) and cache_hint (i64 = 0) + // Cluster G2S carries a default CTA mask. CTA-local G2S has no mask. + // Both forms carry the default cache hint. use pliron::utils::apint::APInt; use std::num::NonZeroUsize; - let i16_type = IntegerType::get(ctx, 16, Signedness::Signed); let i64_type = IntegerType::get(ctx, 64, Signedness::Unsigned); - // Create constant for cta_mask = 0 - let cta_mask_apint = APInt::from_i64(0, NonZeroUsize::new(16).unwrap()); - let cta_mask_attr = pliron::builtin::attributes::IntegerAttr::new(i16_type, cta_mask_apint); - - let cta_mask_raw_op = Operation::new( - ctx, - MirConstantOp::get_concrete_op_info(), - vec![i16_type.to_handle()], - vec![], - vec![], - 0, - ); - cta_mask_raw_op.deref_mut(ctx).set_loc(loc.clone()); - let cta_mask_const = MirConstantOp::new(cta_mask_raw_op); - cta_mask_const.set_attr_value(ctx, cta_mask_attr); - - if let Some(prev) = last_op { - cta_mask_const.get_operation().insert_after(ctx, prev); - } else { - cta_mask_const - .get_operation() - .insert_at_front(block_ptr, ctx); + if !cta { + let i16_type = IntegerType::get(ctx, 16, Signedness::Signed); + let cta_mask_apint = APInt::from_i64(0, NonZeroUsize::new(16).unwrap()); + let cta_mask_attr = pliron::builtin::attributes::IntegerAttr::new(i16_type, cta_mask_apint); + let cta_mask_raw_op = Operation::new( + ctx, + MirConstantOp::get_concrete_op_info(), + vec![i16_type.to_handle()], + vec![], + vec![], + 0, + ); + cta_mask_raw_op.deref_mut(ctx).set_loc(loc.clone()); + let cta_mask_const = MirConstantOp::new(cta_mask_raw_op); + cta_mask_const.set_attr_value(ctx, cta_mask_attr); + if let Some(prev) = last_op { + cta_mask_const.get_operation().insert_after(ctx, prev); + } else { + cta_mask_const + .get_operation() + .insert_at_front(block_ptr, ctx); + } + operands.push(cta_mask_const.get_operation().deref(ctx).get_result(0)); + last_op = Some(cta_mask_const.get_operation()); } - let cta_mask = cta_mask_const.get_operation().deref(ctx).get_result(0); - operands.push(cta_mask); // Create constant for cache_hint = 0 let cache_hint_apint = APInt::from_i64(0, NonZeroUsize::new(64).unwrap()); @@ -183,19 +224,28 @@ pub fn emit_tma_g2s( cache_hint_raw_op.deref_mut(ctx).set_loc(loc.clone()); let cache_hint_const = MirConstantOp::new(cache_hint_raw_op); cache_hint_const.set_attr_value(ctx, cache_hint_attr); - cache_hint_const - .get_operation() - .insert_after(ctx, cta_mask_const.get_operation()); + if let Some(prev) = last_op { + cache_hint_const.get_operation().insert_after(ctx, prev); + } else { + cache_hint_const + .get_operation() + .insert_at_front(block_ptr, ctx); + } let cache_hint = cache_hint_const.get_operation().deref(ctx).get_result(0); operands.push(cache_hint); // Select the appropriate NVVM op based on dimensions let op_id = match dims { + 1 if cta => CpAsyncBulkTensorG2sCtaTile1dOp::get_concrete_op_info(), 1 => CpAsyncBulkTensorG2sTile1dOp::get_concrete_op_info(), + 2 if cta => CpAsyncBulkTensorG2sCtaTile2dOp::get_concrete_op_info(), 2 => CpAsyncBulkTensorG2sTile2dOp::get_concrete_op_info(), + 3 if cta => CpAsyncBulkTensorG2sCtaTile3dOp::get_concrete_op_info(), 3 => CpAsyncBulkTensorG2sTile3dOp::get_concrete_op_info(), + 4 if cta => CpAsyncBulkTensorG2sCtaTile4dOp::get_concrete_op_info(), 4 => CpAsyncBulkTensorG2sTile4dOp::get_concrete_op_info(), + 5 if cta => CpAsyncBulkTensorG2sCtaTile5dOp::get_concrete_op_info(), 5 => CpAsyncBulkTensorG2sTile5dOp::get_concrete_op_info(), _ => { return input_err!( @@ -213,7 +263,7 @@ pub fn emit_tma_g2s( ctx, op_id, vec![], // No results - operands, // dst, barrier, tensor_map, coords..., cta_mask, cache_hint + operands, // dst, barrier, tensor_map, coords..., optional cta_mask, cache_hint vec![], 0, ); diff --git a/cuda-oxide/crates/mir-importer/src/translator/types.rs b/cuda-oxide/crates/mir-importer/src/translator/types.rs index 0b9c79aac..6e693acfa 100644 --- a/cuda-oxide/crates/mir-importer/src/translator/types.rs +++ b/cuda-oxide/crates/mir-importer/src/translator/types.rs @@ -364,6 +364,14 @@ fn translate_pointer_like( let elem = shared_array_element_type(ctx, &substs, "SharedArray")?; Ok(facts::mint_shared_ptr_type(ctx, elem, origin).into()) } + rustc_public::ty::TyKind::RigidTy(rustc_public::ty::RigidTy::Adt(adt_def, substs)) + if is_cuda_device_adt(&adt_def, "SharedSlot") => + { + // `*mut SharedSlot` is the field of `SharedPtr`: a pointer + // to a `T` in CTA-shared memory, by its type. + let elem = shared_array_element_type(ctx, &substs, "SharedSlot")?; + Ok(facts::mint_shared_ptr_type(ctx, elem, origin).into()) + } rustc_public::ty::TyKind::RigidTy(rustc_public::ty::RigidTy::Adt(adt_def, _substs)) if is_cuda_device_adt(&adt_def, "Barrier") => { @@ -718,6 +726,14 @@ pub fn translate_type( // ThreadIndex is a newtype around usize - translate to usize // The type safety is enforced at the Rust level, not the IR level Ok(get_usize_type(ctx).into()) + } else if is_cuda_device_adt(&adt_def, "SharedPtr") { + // `SharedPtr { ptr: *mut SharedSlot }` is a scalar-lowered + // newtype: the value is its field, a CTA-shared (`addrspace(3)`) + // pointer to `T`. Params, returns, locals and fields of type + // `SharedPtr` are therefore shared pointers by type, across + // function boundaries. + let elem = shared_array_element_type(ctx, &substs, "SharedPtr")?; + Ok(facts::mint_shared_ptr_type(ctx, elem, facts::abi_shared_ptr_field()).into()) } else if is_cuda_device_adt(&adt_def, "RangeToken") { Ok(dialect_iket::types::IketRangeTokenType::get(ctx).into()) } else if is_cuda_device_adt(&adt_def, "SharedArray") { diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/active_mask.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/active_mask.rs index 30c877667..2ba93f5ee 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/active_mask.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/active_mask.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `active_mask` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/clc.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/clc.rs index 8fa858647..9b81219db 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/clc.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/clc.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `clc` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cluster_barrier.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cluster_barrier.rs index 4e2734e8d..b72d3d677 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cluster_barrier.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cluster_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `cluster_barrier` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cluster_memory.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cluster_memory.rs index 2829f83d4..eee115f3c 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cluster_memory.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cluster_memory.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `cluster_memory` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cp_async.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cp_async.rs index 2bc5b8882..d336d7570 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cp_async.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/cp_async.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `cp_async` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/debug_control.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/debug_control.rs index 5813b36f7..ab0700d55 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/debug_control.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/debug_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `debug_control` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/dotprod.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/dotprod.rs index 264a05534..e613145a4 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/dotprod.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/dotprod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `dotprod` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/elect.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/elect.rs index efc66dd8b..018ae5b5c 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/elect.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/elect.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `elect` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/execution_control.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/execution_control.rs index 73face4f7..ebf4ddb3a 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/execution_control.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/execution_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `execution_control` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/extended_minmax.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/extended_minmax.rs index f128e318b..1845fb6e4 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/extended_minmax.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/extended_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `extended_minmax` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/integer_minmax.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/integer_minmax.rs index a35c8be28..fe90e07d4 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/integer_minmax.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/integer_minmax.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `integer_minmax` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/ldmatrix.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/ldmatrix.rs index 580155604..e6b93926c 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/ldmatrix.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/ldmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `ldmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mbarrier_basic.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mbarrier_basic.rs index f54e78bfa..5bbc59246 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mbarrier_basic.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mbarrier_basic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `mbarrier_basic` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mbarrier_extended.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mbarrier_extended.rs index 916f51a07..76cf328ca 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mbarrier_extended.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mbarrier_extended.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `mbarrier_extended` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mod.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mod.rs index bfb759ffa..8a8e14904 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mod.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/mod.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces for admitted CUDA intrinsic families. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/movmatrix.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/movmatrix.rs index 49e86c6d9..1a3bc2b87 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/movmatrix.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/movmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `movmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_alu.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_alu.rs index a86511b63..82c5b85ff 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_alu.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_alu.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `packed_alu` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_atomic.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_atomic.rs index 87d5e3e38..5bc05bfcd 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_atomic.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_atomic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `packed_atomic` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_conversion.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_conversion.rs index 5f778e31c..94b3ebd97 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_conversion.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/packed_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `packed_conversion` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/prmt.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/prmt.rs index 8e843de47..1ffac3ad6 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/prmt.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/prmt.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `prmt` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/redux.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/redux.rs index e8106051b..bf6e7322c 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/redux.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/redux.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `redux` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/register_mma.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/register_mma.rs index 719f6e4c0..16d330c5a 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/register_mma.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/register_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `register_mma` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_arithmetic.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_arithmetic.rs index 8fe82b112..284983668 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_arithmetic.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_arithmetic.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `scalar_arithmetic` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_conversion.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_conversion.rs index 278f2e00d..af234e64e 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_conversion.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_conversion.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `scalar_conversion` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_math.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_math.rs index eeda1b07f..bbb32996b 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_math.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/scalar_math.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `scalar_math` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sparse_mma.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sparse_mma.rs index 300f1ed21..7bbeff63d 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sparse_mma.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sparse_mma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `sparse_mma` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sreg.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sreg.rs index 1989dcc8b..6f6fb05c6 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sreg.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sreg.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `sreg` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/stmatrix.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/stmatrix.rs index 7b6f25b87..da17d4e5f 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/stmatrix.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/stmatrix.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `stmatrix` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sync.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sync.rs index f6bea5c03..4f7cd0813 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sync.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/sync.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `sync` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/tcgen05.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/tcgen05.rs index cdc12343b..5c47b505a 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/tcgen05.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/tcgen05.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `tcgen05` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/tma.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/tma.rs index d7306ca5b..a34e31378 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/tma.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/tma.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `tma` intrinsics. @@ -12,9 +12,9 @@ use crate::conversion_interface::MirToLlvmConversion; use crate::convert::intrinsics::{ tma::PrefetchTileConfig, tma::ReduceConfig, tma::convert_control, tma::convert_g2s, - tma::convert_g2s_multicast_cg2, tma::convert_prefetch_tensormap, tma::convert_prefetch_tile, - tma::convert_reduce_s2g, tma::convert_s2g, tma::convert_tensormap_fence, - tma::convert_tensormap_replace, + tma::convert_g2s_cta, tma::convert_g2s_multicast_cg2, tma::convert_prefetch_tensormap, + tma::convert_prefetch_tile, tma::convert_reduce_s2g, tma::convert_s2g, + tma::convert_tensormap_fence, tma::convert_tensormap_replace, }; use dialect_nvvm::ops::{ CpAsyncBulkCommitGroupOp, CpAsyncBulkPrefetchTensor1dL2CacheHintOp, @@ -23,7 +23,9 @@ use dialect_nvvm::ops::{ CpAsyncBulkPrefetchTensor3dL2Op, CpAsyncBulkPrefetchTensor4dL2CacheHintOp, CpAsyncBulkPrefetchTensor4dL2Op, CpAsyncBulkPrefetchTensor5dL2CacheHintOp, CpAsyncBulkPrefetchTensor5dL2Op, CpAsyncBulkPrefetchTensorGather4TwoDimensionalL2CacheHintOp, - CpAsyncBulkPrefetchTensorGather4TwoDimensionalL2Op, CpAsyncBulkTensorG2sTile1dOp, + CpAsyncBulkPrefetchTensorGather4TwoDimensionalL2Op, CpAsyncBulkTensorG2sCtaTile1dOp, + CpAsyncBulkTensorG2sCtaTile2dOp, CpAsyncBulkTensorG2sCtaTile3dOp, + CpAsyncBulkTensorG2sCtaTile4dOp, CpAsyncBulkTensorG2sCtaTile5dOp, CpAsyncBulkTensorG2sTile1dOp, CpAsyncBulkTensorG2sTile2dMulticastCg2Op, CpAsyncBulkTensorG2sTile2dMulticastOp, CpAsyncBulkTensorG2sTile2dOp, CpAsyncBulkTensorG2sTile3dOp, CpAsyncBulkTensorG2sTile4dOp, CpAsyncBulkTensorG2sTile5dOp, CpAsyncBulkTensorReduceAddIm2col3dOp, @@ -383,6 +385,18 @@ impl MirToLlvmConversion for CpAsyncBulkTensorG2sTile1dOp { } } +#[op_interface_impl] +impl MirToLlvmConversion for CpAsyncBulkTensorG2sCtaTile1dOp { + fn convert( + &self, + ctx: &mut Context, + rewriter: &mut DialectConversionRewriter, + operands_info: &OperandsInfo, + ) -> Result<()> { + convert_g2s_cta(ctx, rewriter, self.get_operation(), operands_info, 1) + } +} + #[op_interface_impl] impl MirToLlvmConversion for CpAsyncBulkTensorS2gTile1dOp { fn convert( @@ -407,6 +421,18 @@ impl MirToLlvmConversion for CpAsyncBulkTensorG2sTile2dOp { } } +#[op_interface_impl] +impl MirToLlvmConversion for CpAsyncBulkTensorG2sCtaTile2dOp { + fn convert( + &self, + ctx: &mut Context, + rewriter: &mut DialectConversionRewriter, + operands_info: &OperandsInfo, + ) -> Result<()> { + convert_g2s_cta(ctx, rewriter, self.get_operation(), operands_info, 2) + } +} + #[op_interface_impl] impl MirToLlvmConversion for CpAsyncBulkTensorG2sTile2dMulticastOp { fn convert( @@ -455,6 +481,18 @@ impl MirToLlvmConversion for CpAsyncBulkTensorG2sTile3dOp { } } +#[op_interface_impl] +impl MirToLlvmConversion for CpAsyncBulkTensorG2sCtaTile3dOp { + fn convert( + &self, + ctx: &mut Context, + rewriter: &mut DialectConversionRewriter, + operands_info: &OperandsInfo, + ) -> Result<()> { + convert_g2s_cta(ctx, rewriter, self.get_operation(), operands_info, 3) + } +} + #[op_interface_impl] impl MirToLlvmConversion for CpAsyncBulkTensorS2gTile3dOp { fn convert( @@ -479,6 +517,18 @@ impl MirToLlvmConversion for CpAsyncBulkTensorG2sTile4dOp { } } +#[op_interface_impl] +impl MirToLlvmConversion for CpAsyncBulkTensorG2sCtaTile4dOp { + fn convert( + &self, + ctx: &mut Context, + rewriter: &mut DialectConversionRewriter, + operands_info: &OperandsInfo, + ) -> Result<()> { + convert_g2s_cta(ctx, rewriter, self.get_operation(), operands_info, 4) + } +} + #[op_interface_impl] impl MirToLlvmConversion for CpAsyncBulkTensorS2gTile4dOp { fn convert( @@ -503,6 +553,18 @@ impl MirToLlvmConversion for CpAsyncBulkTensorG2sTile5dOp { } } +#[op_interface_impl] +impl MirToLlvmConversion for CpAsyncBulkTensorG2sCtaTile5dOp { + fn convert( + &self, + ctx: &mut Context, + rewriter: &mut DialectConversionRewriter, + operands_info: &OperandsInfo, + ) -> Result<()> { + convert_g2s_cta(ctx, rewriter, self.get_operation(), operands_info, 5) + } +} + #[op_interface_impl] impl MirToLlvmConversion for CpAsyncBulkTensorS2gTile5dOp { fn convert( diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/vote.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/vote.rs index 8ec94c0b3..219add438 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/vote.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/vote.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `vote` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_barrier.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_barrier.rs index cfebdc112..43a0524d5 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_barrier.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_barrier.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `warp_barrier` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_match.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_match.rs index 3516d587b..7c6eb0c56 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_match.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_match.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `warp_match` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_shuffle.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_shuffle.rs index 2a8c876e7..05ab52282 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_shuffle.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/warp_shuffle.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `warp_shuffle` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/wgmma_control.rs b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/wgmma_control.rs index 8fb47633d..dd8526e8d 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/wgmma_control.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/generated_intrinsics/wgmma_control.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated conversion interfaces: `wgmma_control` intrinsics. diff --git a/cuda-oxide/crates/mir-lower/src/convert/intrinsics/tma.rs b/cuda-oxide/crates/mir-lower/src/convert/intrinsics/tma.rs index 9c5fcae46..66a51c81e 100644 --- a/cuda-oxide/crates/mir-lower/src/convert/intrinsics/tma.rs +++ b/cuda-oxide/crates/mir-lower/src/convert/intrinsics/tma.rs @@ -58,6 +58,166 @@ fn g2s_inline_asm(dims: usize, multicast: bool, cta_group: i32) -> (String, Stri (template, constraints.join(",")) } +fn g2s_cta_inline_asm(dims: usize) -> (String, String) { + let coordinates = (0..dims) + .map(|index| format!("${}", 3 + index)) + .collect::>() + .join(", "); + let template = format!( + "cp.async.bulk.tensor.{dims}d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$0], [$2, {{{coordinates}}}], [$1];" + ); + let mut constraints = vec!["r", "r", "l"]; + constraints.extend(std::iter::repeat_n("r", dims)); + constraints.push("~{memory}"); + (template, constraints.join(",")) +} + +/// Address space of a lowered pointer operand, or `None` if it is not a +/// pointer. +fn pointer_address_space(ctx: &Context, value: pliron::value::Value) -> Option { + value + .get_type(ctx) + .deref(ctx) + .downcast_ref::() + .map(|pointer| pointer.address_space()) +} + +/// Convert a TMA G2S operation whose destination and barrier belong to the +/// issuing CTA rather than to cluster shared memory. +/// +/// The destination is shared by type: the device wrapper takes a +/// `SharedPtr`, which lowers to an `addrspace(3)` pointer, while the raw +/// ABI entry still passes a generic pointer under the caller's promise of +/// CTA locality. The barrier is a caller promise on both layers, as for every +/// mbarrier operation. Generic and shared pointers are converted to the +/// shared window. A pointer known to be cluster shared (`addrspace(7)`, for +/// example from `map_shared_rank`) may name another CTA's shared memory, so +/// it is rejected rather than reinterpreted as a local offset; such copies +/// use the cluster form. The check sees only uncast operands: a pointer cast +/// or `SharedPtr::from_raw` applied to a cluster pointer is the caller's +/// unsafe contract. +pub(crate) fn convert_g2s_cta( + ctx: &mut Context, + rewriter: &mut DialectConversionRewriter, + op: Ptr, + _operands_info: &OperandsInfo, + dims: usize, +) -> Result<()> { + let operands: Vec<_> = op.deref(ctx).operands().collect(); + let expected_operands = 3 + dims + 1; + if operands.len() != expected_operands { + return pliron::input_err_noloc!( + "TMA CTA G2S {}D requires {} operands, got {}", + dims, + expected_operands, + operands.len() + ); + } + for (index, role) in [(0, "destination"), (1, "barrier")] { + match pointer_address_space(ctx, operands[index]) { + Some(0 | 3) => {} + Some(7) => { + return pliron::input_err!( + op.deref(ctx).loc(), + "cp_async_bulk_tensor_{}d_g2s_cta requires a CTA-local shared {}, \ + but it is a cluster shared memory pointer (addrspace 7), which may \ + name another CTA's shared memory; use cp_async_bulk_tensor_{}d_g2s \ + for cluster destinations", + dims, + role, + dims + ); + } + Some(space) => { + return pliron::input_err!( + op.deref(ctx).loc(), + "cp_async_bulk_tensor_{}d_g2s_cta requires a shared memory {}, \ + but it is in address space {}", + dims, + role, + space + ); + } + None => { + return pliron::input_err!( + op.deref(ctx).loc(), + "cp_async_bulk_tensor_{}d_g2s_cta {} must be a pointer", + dims, + role + ); + } + } + } + + let dst_shared = cast_to_shared_addrspace(ctx, rewriter, operands[0]); + let barrier_shared = cast_to_shared_addrspace(ctx, rewriter, operands[1]); + let void_ty = llvm_types::VoidType::get(ctx); + + if context::lowering_options(ctx).intrinsic_backend == IntrinsicBackend::LibNvvm { + // Legacy NVVM IR has no typed CTA-destination intrinsic. Pass the + // 32-bit shared-window addresses to the exact PTX instruction. + let i32_ty = IntegerType::get(ctx, 32, Signedness::Signless); + let dst_address = llvm::PtrToIntOp::new(ctx, dst_shared, i32_ty.into()); + rewriter.insert_operation(ctx, dst_address.get_operation()); + let barrier_address = llvm::PtrToIntOp::new(ctx, barrier_shared, i32_ty.into()); + rewriter.insert_operation(ctx, barrier_address.get_operation()); + let mut inputs = vec![ + dst_address.get_operation().deref(ctx).get_result(0), + barrier_address.get_operation().deref(ctx).get_result(0), + operands[2], + ]; + inputs.extend(operands[3..3 + dims].iter().copied()); + let (template, constraints) = g2s_cta_inline_asm(dims); + inline_asm_convergent( + ctx, + rewriter, + op, + void_ty.into(), + inputs, + &template, + &constraints, + ); + rewriter.erase_operation(ctx, op); + return Ok(()); + } + + // LLVM NVPTX: the typed intrinsic the catalog probes record. + let i32_ty = IntegerType::get(ctx, 32, Signedness::Signless); + let i64_ty = IntegerType::get(ctx, 64, Signedness::Signless); + let i1_ty = IntegerType::get(ctx, 1, Signedness::Signless); + let smem_ptr_ty = llvm_types::PointerType::get(ctx, 3); + let generic_ptr_ty = llvm_types::PointerType::get(ctx, 0); + let mut arg_types: Vec = vec![ + smem_ptr_ty.into(), + smem_ptr_ty.into(), + generic_ptr_ty.into(), + ]; + for _ in 0..dims { + arg_types.push(i32_ty.into()); + } + arg_types.push(i64_ty.into()); // cache_hint + arg_types.push(i1_ty.into()); // use_cache_hint + + let intrinsic_name = format!("llvm_nvvm_cp_async_bulk_tensor_g2s_cta_tile_{}d", dims); + let func_ty = llvm_types::FuncType::get(ctx, void_ty.into(), arg_types, false); + let parent_block = op.deref(ctx).get_parent_block().unwrap(); + helpers::ensure_intrinsic_declared(ctx, parent_block, &intrinsic_name, func_ty) + .map_err(|e| pliron::input_error_noloc!("{}", e))?; + + let mut call_args = vec![dst_shared, barrier_shared]; + call_args.extend(operands[2..].iter().copied()); + let use_cache_hint = create_i1_const(ctx, rewriter, false); + call_args.push(use_cache_hint); + + let sym_name: pliron::identifier::Identifier = intrinsic_name.as_str().try_into().unwrap(); + let callee = CallOpCallable::Direct(sym_name); + let llvm_call = llvm::CallOp::new(ctx, callee, func_ty, call_args); + crate::convert::preserve_location(ctx, op, llvm_call.get_operation()); + rewriter.insert_operation(ctx, llvm_call.get_operation()); + rewriter.erase_operation(ctx, op); + Ok(()) +} + fn convert_g2s_impl( ctx: &mut Context, rewriter: &mut DialectConversionRewriter, @@ -746,7 +906,7 @@ pub(crate) fn convert_control( #[cfg(test)] mod tests { - use super::{g2s_inline_asm, reduce_inline_asm, s2g_inline_asm}; + use super::{g2s_cta_inline_asm, g2s_inline_asm, reduce_inline_asm, s2g_inline_asm}; #[test] fn inline_tma_templates_keep_exact_ptx_shapes() { @@ -764,6 +924,13 @@ mod tests { "l,l,l,r,r,h,~{memory}".into(), ) ); + assert_eq!( + g2s_cta_inline_asm(2), + ( + "cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$0], [$2, {$3, $4}], [$1];".into(), + "r,r,l,r,r,~{memory}".into(), + ) + ); assert_eq!( s2g_inline_asm(5), ( diff --git a/cuda-oxide/crates/mir-lower/tests/lowering_test/tma.rs b/cuda-oxide/crates/mir-lower/tests/lowering_test/tma.rs index 7cdf3f6d8..1215c19cd 100644 --- a/cuda-oxide/crates/mir-lower/tests/lowering_test/tma.rs +++ b/cuda-oxide/crates/mir-lower/tests/lowering_test/tma.rs @@ -171,3 +171,147 @@ fn tma_modern_libnvvm_preserves_existing_cluster_pointer_semantics() -> Result<( assert!(!ir.contains("cvta.to.shared::cta"), "{ir}"); Ok(()) } + +/// Lowers the five CTA-local G2S forms with the given destination and +/// barrier address spaces. +fn lower_g2s_cta_forms( + destination_space: u32, + barrier_space: u32, + backend: mir_lower::IntrinsicBackend, +) -> Result<(Context, Ptr), anyhow::Error> { + let mut ctx = make_test_ctx(); + let byte = IntegerType::get(&ctx, 8, Signedness::Signless).into(); + let i32_ty = IntegerType::get(&ctx, 32, Signedness::Signless).into(); + let i64_ty = IntegerType::get(&ctx, 64, Signedness::Signless).into(); + let destination = MirPtrType::get(&mut ctx, byte, true, destination_space).into(); + let barrier = MirPtrType::get(&mut ctx, byte, true, barrier_space).into(); + let pointer = MirPtrType::get_generic(&mut ctx, byte, true).into(); + let (module, entry) = build_test_kernel( + &mut ctx, + vec![ + destination, + barrier, + pointer, + i32_ty, + i32_ty, + i32_ty, + i32_ty, + i32_ty, + i64_ty, + ], + ); + let args: Vec<_> = (0..9) + .map(|index| entry.deref(&ctx).get_argument(index)) + .collect(); + for (info, dimensions) in [ + ( + nvvm::CpAsyncBulkTensorG2sCtaTile1dOp::get_concrete_op_info(), + 1, + ), + ( + nvvm::CpAsyncBulkTensorG2sCtaTile2dOp::get_concrete_op_info(), + 2, + ), + ( + nvvm::CpAsyncBulkTensorG2sCtaTile3dOp::get_concrete_op_info(), + 3, + ), + ( + nvvm::CpAsyncBulkTensorG2sCtaTile4dOp::get_concrete_op_info(), + 4, + ), + ( + nvvm::CpAsyncBulkTensorG2sCtaTile5dOp::get_concrete_op_info(), + 5, + ), + ] { + let mut operands = args[..3 + dimensions].to_vec(); + operands.push(args[8]); + Operation::new(&mut ctx, info, vec![], operands, vec![], 0).insert_at_back(entry, &ctx); + } + append_return(&mut ctx, entry); + mir_lower::lower_mir_to_llvm_with_options( + &mut ctx, + module, + mir_lower::LoweringOptions { + intrinsic_backend: backend, + ..Default::default() + }, + ) + .map_err(|error| anyhow::anyhow!("{error}"))?; + Ok((ctx, module)) +} + +#[test] +fn tma_cta_llvm_nvptx_uses_typed_shared_cta_intrinsics() -> Result<(), anyhow::Error> { + for (destination_space, barrier_space) in [(0, 0), (3, 3), (0, 3), (3, 0)] { + let (ctx, module) = lower_g2s_cta_forms( + destination_space, + barrier_space, + mir_lower::IntrinsicBackend::LlvmNvptx, + )?; + let module = Operation::get_op::(module, &ctx).unwrap(); + let ir = llvm_export::export::export_module_to_string(&ctx, &module).unwrap(); + for dimensions in 1..=5 { + let call = format!( + "call void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.{dimensions}d(ptr addrspace(3)" + ); + assert_eq!(ir.matches(&call).count(), 1, "{ir}"); + } + assert!(!ir.contains("addrspace(7)"), "{ir}"); + assert!(!ir.contains("g2s.tile."), "{ir}"); + assert!(!ir.contains("asm sideeffect"), "{ir}"); + } + Ok(()) +} + +#[test] +fn tma_cta_libnvvm_emits_exact_shared_cta_ptx() -> Result<(), anyhow::Error> { + for (destination_space, barrier_space) in [(0, 0), (3, 3)] { + let (ctx, module) = lower_g2s_cta_forms( + destination_space, + barrier_space, + mir_lower::IntrinsicBackend::LibNvvm, + )?; + let module = Operation::get_op::(module, &ctx).unwrap(); + for dialect in [NvvmIrDialect::LegacyLlvm7, NvvmIrDialect::Modern] { + let ir = + export_module_to_string_with_config(&ctx, &module, &NvvmExportConfig::new(dialect)) + .unwrap(); + for dimensions in 1..=5 { + let instruction = format!( + "cp.async.bulk.tensor.{dimensions}d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$0], [$2," + ); + assert_eq!(ir.matches(&instruction).count(), 1, "{ir}"); + } + assert!(!ir.contains("addrspace(7)"), "{ir}"); + assert!(!ir.contains("shared::cluster"), "{ir}"); + } + } + Ok(()) +} + +#[test] +fn tma_cta_rejects_cluster_shared_operands() { + for backend in [ + mir_lower::IntrinsicBackend::LlvmNvptx, + mir_lower::IntrinsicBackend::LibNvvm, + ] { + for (destination_space, barrier_space, role) in [ + (7, 3, "destination"), + (3, 7, "barrier"), + (7, 0, "destination"), + ] { + let error = lower_g2s_cta_forms(destination_space, barrier_space, backend) + .err() + .unwrap_or_else(|| panic!("an addrspace(7) {role} lowered with {backend:?}")) + .to_string(); + assert!( + error.contains("g2s_cta requires a CTA-local shared") + && error.contains(role) + && error.contains("addrspace 7"), + "{error}" + ); + } + } +} diff --git a/cuda-oxide/crates/rustc-codegen-cuda/STATUS.md b/cuda-oxide/crates/rustc-codegen-cuda/STATUS.md index f42a90129..94b339cbd 100644 --- a/cuda-oxide/crates/rustc-codegen-cuda/STATUS.md +++ b/cuda-oxide/crates/rustc-codegen-cuda/STATUS.md @@ -25,6 +25,7 @@ Run `cuda-oxide/scripts/check-error-example-status.sh` to verify both are in syn | `error_host_arch_intrinsic` | diagnostics-fixture | `core::arch::` intrinsic reached | | `error_host_target_feature` | diagnostics-fixture | `#[target_feature]` fn reached | | `error_kernel_shared_param` | diagnostics-fixture | AS3/AS5 pointer as kernel parameter | +| `error_kernel_shared_ptr_param` | diagnostics-fixture | AS3 pointer inside a kernel parameter struct | | `error_missing_device_attr` | diagnostics-fixture | `thread::index_*` stub (#76) | | `error_set_discriminant_uninhabited` | diagnostics-fixture | Invalid enum variant selection | | `error_union_shared_pointer_addr` | diagnostics-fixture | `&mut`/in-place write to a shared-pointer union field | diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/Cargo.lock b/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/Cargo.lock new file mode 100644 index 000000000..c46db5e9e --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/Cargo.lock @@ -0,0 +1,753 @@ +# This file is automatically @generated by Cargo. +# It is not intended for manual editing. +version = 4 + +[[package]] +name = "aho-corasick" +version = "1.1.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ddd31a130427c27518df266943a5308ed92d4b226cc639f5a8f1002816174301" +dependencies = [ + "memchr", +] + +[[package]] +name = "anyhow" +version = "1.0.102" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7f202df86484c868dbad7eaa557ef785d5c66295e41b460ef922eca0723b842c" + +[[package]] +name = "bindgen" +version = "0.69.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "271383c67ccabffb7381723dea0672a673f292304fcb45c01cc648c7a8d58088" +dependencies = [ + "bitflags", + "cexpr", + "clang-sys", + "itertools", + "lazy_static", + "lazycell", + "log", + "prettyplease", + "proc-macro2", + "quote", + "regex", + "rustc-hash", + "shlex", + "syn", + "which", +] + +[[package]] +name = "bitflags" +version = "2.13.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b4388bee8683e3d04af747c73422af53102d2bd24d9eadb6cbc100baef4b43f8" + +[[package]] +name = "block-buffer" +version = "0.10.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3078c7629b62d3f0439517fa394996acacc5cbc91c5a20d8c658e77abd503a71" +dependencies = [ + "generic-array", +] + +[[package]] +name = "cexpr" +version = "0.6.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6fac387a98bb7c37292057cffc56d62ecb629900026402633ae9160df93a8766" +dependencies = [ + "nom", +] + +[[package]] +name = "cfg-if" +version = "1.0.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9330f8b2ff13f34540b44e946ef35111825727b38d33286ef986142615121801" + +[[package]] +name = "clang-sys" +version = "1.8.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0b023947811758c97c59bf9d1c188fd619ad4718dcaa767947df1cadb14f39f4" +dependencies = [ + "glob", + "libc", + "libloading", +] + +[[package]] +name = "cpufeatures" +version = "0.2.17" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "59ed5838eebb26a2bb2e58f6d5b5316989ae9d08bab10e0e6d103e656d1b0280" +dependencies = [ + "libc", +] + +[[package]] +name = "crc32fast" +version = "1.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9481c1c90cbf2ac953f07c8d4a58aa3945c425b7185c9154d67a65e4230da511" +dependencies = [ + "cfg-if", +] + +[[package]] +name = "crunchy" +version = "0.2.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "460fbee9c2c2f33933d720630a6a0bac33ba7053db5344fac858d4b8952d77d5" + +[[package]] +name = "crypto-common" +version = "0.1.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "78c8292055d1c1df0cce5d180393dc8cce0abec0a7102adb6c7b1eef6016d60a" +dependencies = [ + "generic-array", + "typenum", +] + +[[package]] +name = "cuda-artifact-finalizer" +version = "0.2.1" +dependencies = [ + "libnvvm-sys", + "nvjitlink-sys", + "serde", + "sha2", + "thiserror", +] + +[[package]] +name = "cuda-bindings" +version = "0.4.0" +dependencies = [ + "bindgen", + "libloading", + "prettyplease", + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "cuda-core" +version = "0.4.0" +dependencies = [ + "anyhow", + "cuda-bindings", + "cuda-core-derive", + "half", + "oxide-artifacts", +] + +[[package]] +name = "cuda-core-derive" +version = "0.4.0" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "cuda-device" +version = "0.2.1" +dependencies = [ + "cuda-macros", +] + +[[package]] +name = "cuda-host" +version = "0.2.1" +dependencies = [ + "cuda-artifact-finalizer", + "cuda-core", + "cuda-macros", + "half", + "libc", + "oxide-artifacts", + "ptx-parse", + "sha2", + "thiserror", +] + +[[package]] +name = "cuda-macros" +version = "0.2.1" +dependencies = [ + "oxide-tile-export", + "proc-macro2", + "quote", + "reserved-oxide-symbols", + "syn", +] + +[[package]] +name = "cuda-target-spec" +version = "0.1.0" + +[[package]] +name = "digest" +version = "0.10.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9ed9a281f7bc9b7576e61468ba615a66a5c8cfdff42420a70aa82701a3b1e292" +dependencies = [ + "block-buffer", + "crypto-common", +] + +[[package]] +name = "either" +version = "1.16.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "91622ff5e7162018101f2fea40d6ebf4a78bbe5a49736a2020649edf9693679e" + +[[package]] +name = "equivalent" +version = "1.0.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "877a4ace8713b0bcf2a4e7eec82529c029f1d0619886d18145fea96c3ffe5c0f" + +[[package]] +name = "errno" +version = "0.3.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "39cab71617ae0d63f51a36d69f866391735b51691dbda63cf6f96d042b63efeb" +dependencies = [ + "libc", + "windows-sys 0.61.2", +] + +[[package]] +name = "error_kernel_shared_ptr_param" +version = "0.1.0" +dependencies = [ + "cuda-core", + "cuda-device", + "cuda-host", +] + +[[package]] +name = "foldhash" +version = "0.1.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d9c4f5dac5e15c24eb999c26181a6ca40b39fe946cbe4c263c7209467bc83af2" + +[[package]] +name = "generic-array" +version = "0.14.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "85649ca51fd72272d7821adaf274ad91c288277713d9c18820d8499a7ff69e9a" +dependencies = [ + "typenum", + "version_check", +] + +[[package]] +name = "glob" +version = "0.3.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0cc23270f6e1808e30a928bdc84dea0b9b4136a8bc82338574f23baf47bbd280" + +[[package]] +name = "half" +version = "2.7.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6ea2d84b969582b4b1864a92dc5d27cd2b77b622a8d79306834f1be5ba20d84b" +dependencies = [ + "cfg-if", + "crunchy", + "zerocopy", +] + +[[package]] +name = "hashbrown" +version = "0.15.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9229cfe53dfd69f0609a49f65461bd93001ea1ef889cd5529dd176593f5338a1" +dependencies = [ + "foldhash", +] + +[[package]] +name = "hashbrown" +version = "0.17.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ed5909b6e89a2db4456e54cd5f673791d7eca6732202bbf2a9cc504fe2f9b84a" + +[[package]] +name = "home" +version = "0.5.12" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cc627f471c528ff0c4a49e1d5e60450c8f6461dd6d10ba9dcd3a61d3dff7728d" +dependencies = [ + "windows-sys 0.61.2", +] + +[[package]] +name = "indexmap" +version = "2.14.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d466e9454f08e4a911e14806c24e16fba1b4c121d1ea474396f396069cf949d9" +dependencies = [ + "equivalent", + "hashbrown 0.17.1", +] + +[[package]] +name = "itertools" +version = "0.12.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ba291022dbbd398a455acf126c1e341954079855bc60dfdda641363bd6922569" +dependencies = [ + "either", +] + +[[package]] +name = "itoa" +version = "1.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682" + +[[package]] +name = "lazy_static" +version = "1.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe" + +[[package]] +name = "lazycell" +version = "1.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55" + +[[package]] +name = "libc" +version = "0.2.186" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "68ab91017fe16c622486840e4c83c9a37afeff978bd239b5293d61ece587de66" + +[[package]] +name = "libloading" +version = "0.8.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d7c4b02199fee7c5d21a5ae7d8cfa79a6ef5bb2fc834d6e9058e89c825efdc55" +dependencies = [ + "cfg-if", + "windows-link", +] + +[[package]] +name = "libnvvm-sys" +version = "0.2.1" +dependencies = [ + "cuda-target-spec", + "libloading", + "thiserror", +] + +[[package]] +name = "linux-raw-sys" +version = "0.4.15" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d26c52dbd32dccf2d10cac7725f8eae5296885fb5703b261f7d0a0739ec807ab" + +[[package]] +name = "log" +version = "0.4.29" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5e5032e24019045c762d3c0f28f5b6b8bbf38563a65908389bf7978758920897" + +[[package]] +name = "memchr" +version = "2.8.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cf8baf1c55e62ffcace7a9f06f4bd9cd3f0c4beb022d3b367256b91b87513d98" + +[[package]] +name = "minimal-lexical" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a" + +[[package]] +name = "nom" +version = "7.1.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a" +dependencies = [ + "memchr", + "minimal-lexical", +] + +[[package]] +name = "nvjitlink-sys" +version = "0.2.1" +dependencies = [ + "libloading", + "thiserror", +] + +[[package]] +name = "object" +version = "0.36.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "62948e14d923ea95ea2c7c86c71013138b66525b86bdc08d2dcc262bdb497b87" +dependencies = [ + "crc32fast", + "hashbrown 0.15.5", + "indexmap", + "memchr", +] + +[[package]] +name = "once_cell" +version = "1.21.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9f7c3e4beb33f85d45ae3e3a1792185706c8e16d043238c593331cc7cd313b50" + +[[package]] +name = "oxide-artifacts" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "74d9a941c0c43d2dceaf25f5c9ca4c9f9cceb28dfa5e4e3cfd14e476f003e7c9" +dependencies = [ + "object", +] + +[[package]] +name = "oxide-tile-export" +version = "0.2.1" +dependencies = [ + "serde", + "serde_json", + "thiserror", +] + +[[package]] +name = "prettyplease" +version = "0.2.37" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "479ca8adacdd7ce8f1fb39ce9ecccbfe93a3f1344b3d0d97f20bc0196208f62b" +dependencies = [ + "proc-macro2", + "syn", +] + +[[package]] +name = "proc-macro2" +version = "1.0.106" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8fd00f0bb2e90d81d1044c2b32617f68fcb9fa3bb7640c23e9c748e53fb30934" +dependencies = [ + "unicode-ident", +] + +[[package]] +name = "ptx-parse" +version = "0.1.0" + +[[package]] +name = "quote" +version = "1.0.45" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "41f2619966050689382d2b44f664f4bc593e129785a36d6ee376ddf37259b924" +dependencies = [ + "proc-macro2", +] + +[[package]] +name = "regex" +version = "1.12.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e10754a14b9137dd7b1e3e5b0493cc9171fdd105e0ab477f51b72e7f3ac0e276" +dependencies = [ + "aho-corasick", + "memchr", + "regex-automata", + "regex-syntax", +] + +[[package]] +name = "regex-automata" +version = "0.4.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6e1dd4122fc1595e8162618945476892eefca7b88c52820e74af6262213cae8f" +dependencies = [ + "aho-corasick", + "memchr", + "regex-syntax", +] + +[[package]] +name = "regex-syntax" +version = "0.8.11" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d6f6ff9a378485b298a5286656da665ba74413d36db0979633275d2e708145d4" + +[[package]] +name = "reserved-oxide-symbols" +version = "0.2.1" + +[[package]] +name = "rustc-hash" +version = "1.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2" + +[[package]] +name = "rustix" +version = "0.38.44" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "fdb5bc1ae2baa591800df16c9ca78619bf65c0488b41b96ccec5d11220d8c154" +dependencies = [ + "bitflags", + "errno", + "libc", + "linux-raw-sys", + "windows-sys 0.59.0", +] + +[[package]] +name = "serde" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9a8e94ea7f378bd32cbbd37198a4a91436180c5bb472411e48b5ec2e2124ae9e" +dependencies = [ + "serde_core", + "serde_derive", +] + +[[package]] +name = "serde_core" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "41d385c7d4ca58e59fc732af25c3983b67ac852c1a25000afe1175de458b67ad" +dependencies = [ + "serde_derive", +] + +[[package]] +name = "serde_derive" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d540f220d3187173da220f885ab66608367b6574e925011a9353e4badda91d79" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "serde_json" +version = "1.0.154" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e7e9cc8b1b85264074fbcc02a88680c4096b1e47df8f739dceb03bf482f04bd6" +dependencies = [ + "itoa", + "memchr", + "serde", + "serde_core", + "zmij", +] + +[[package]] +name = "sha2" +version = "0.10.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a7507d819769d01a365ab707794a4084392c824f54a7a6a7862f8c3d0892b283" +dependencies = [ + "cfg-if", + "cpufeatures", + "digest", +] + +[[package]] +name = "shlex" +version = "1.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64" + +[[package]] +name = "syn" +version = "2.0.117" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e665b8803e7b1d2a727f4023456bbbbe74da67099c585258af0ad9c5013b9b99" +dependencies = [ + "proc-macro2", + "quote", + "unicode-ident", +] + +[[package]] +name = "thiserror" +version = "2.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4288b5bcbc7920c07a1149a35cf9590a2aa808e0bc1eafaade0b80947865fbc4" +dependencies = [ + "thiserror-impl", +] + +[[package]] +name = "thiserror-impl" +version = "2.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ebc4ee7f67670e9b64d05fa4253e753e016c6c95ff35b89b7941d6b856dec1d5" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "typenum" +version = "1.20.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b6f5e870be6c3b371b77fe0ee0bafb859fa4964b4404c27de1d380043c4dda20" + +[[package]] +name = "unicode-ident" +version = "1.0.24" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e6e4313cd5fcd3dad5cafa179702e2b244f760991f45397d14d4ebf38247da75" + +[[package]] +name = "version_check" +version = "0.9.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0b928f33d975fc6ad9f86c8f283853ad26bdd5b10b7f1542aa2fa15e2289105a" + +[[package]] +name = "which" +version = "4.4.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "87ba24419a2078cd2b0f2ede2691b6c66d8e47836da3b6db8265ebad47afbfc7" +dependencies = [ + "either", + "home", + "once_cell", + "rustix", +] + +[[package]] +name = "windows-link" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f0805222e57f7521d6a62e36fa9163bc891acd422f971defe97d64e70d0a4fe5" + +[[package]] +name = "windows-sys" +version = "0.59.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1e38bc4d79ed67fd075bcc251a1c39b32a1776bbe92e5bef1f0bf1f8c531853b" +dependencies = [ + "windows-targets", +] + +[[package]] +name = "windows-sys" +version = "0.61.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ae137229bcbd6cdf0f7b80a31df61766145077ddf49416a728b02cb3921ff3fc" +dependencies = [ + "windows-link", +] + +[[package]] +name = "windows-targets" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9b724f72796e036ab90c1021d4780d4d3d648aca59e491e6b98e725b84e99973" +dependencies = [ + "windows_aarch64_gnullvm", + "windows_aarch64_msvc", + "windows_i686_gnu", + "windows_i686_gnullvm", + "windows_i686_msvc", + "windows_x86_64_gnu", + "windows_x86_64_gnullvm", + "windows_x86_64_msvc", +] + +[[package]] +name = "windows_aarch64_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "32a4622180e7a0ec044bb555404c800bc9fd9ec262ec147edd5989ccd0c02cd3" + +[[package]] +name = "windows_aarch64_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "09ec2a7bb152e2252b53fa7803150007879548bc709c039df7627cabbd05d469" + +[[package]] +name = "windows_i686_gnu" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8e9b5ad5ab802e97eb8e295ac6720e509ee4c243f69d781394014ebfe8bbfa0b" + +[[package]] +name = "windows_i686_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0eee52d38c090b3caa76c563b86c3a4bd71ef1a819287c19d586d7334ae8ed66" + +[[package]] +name = "windows_i686_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "240948bc05c5e7c6dabba28bf89d89ffce3e303022809e73deaefe4f6ec56c66" + +[[package]] +name = "windows_x86_64_gnu" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "147a5c80aabfbf0c7d901cb5895d1de30ef2907eb21fbbab29ca94c5b08b1a78" + +[[package]] +name = "windows_x86_64_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "24d5b23dc417412679681396f2b49f3de8c1473deb516bd34410872eff51ed0d" + +[[package]] +name = "windows_x86_64_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "589f6da84c646204747d1270a2a5661ea66ed1cced2631d546fdfb155959f9ec" + +[[package]] +name = "zerocopy" +version = "0.8.52" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ce1022995ff5ff5d841ad7d994facc23098cd40152f2c1d11cd607c6f530653f" +dependencies = [ + "zerocopy-derive", +] + +[[package]] +name = "zerocopy-derive" +version = "0.8.52" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1ae7f38b72ec2a254e2b87ef277cf2cd4fb97cbebf944faa6f33354da0867930" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "zmij" +version = "1.0.23" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "29666d0abbfad1e3dc4dcf6144730dd3a3ab225bbbdac83319345b1b44ccfc1b" diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/Cargo.toml b/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/Cargo.toml new file mode 100644 index 000000000..15fee6ddd --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/Cargo.toml @@ -0,0 +1,13 @@ +[package] +name = "error_kernel_shared_ptr_param" +version = "0.1.0" +edition = "2024" +authors = ["NVIDIA Corporation"] +license = "Apache-2.0" + +[workspace] + +[dependencies] +cuda-core = { path = "../../../../../cuda-core" } +cuda-device = { path = "../../../cuda-device" } +cuda-host = { path = "../../../cuda-host" } diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/README.md b/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/README.md new file mode 100644 index 000000000..0fcfdc5d9 --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/README.md @@ -0,0 +1,32 @@ +# `error_kernel_shared_ptr_param` + +Negative test for a `#[kernel]` whose parameter is a struct that holds a +`SharedPtr`, a pointer that lowers to `addrspace(3)` by its type. + +A kernel receives its parameters in `.param` space, filled by the host at +launch. Shared memory is allocated per block by the device, so the host holds +no address in it to pass. `error_kernel_shared_param` pins the refusal of a +bare shared pointer parameter. This fixture pins the same refusal for a +pointer carried inside an aggregate: before the exporter looked through struct +fields and array elements, `Args { p: SharedPtr, n: u32 }` compiled, the +entry declared a 16-byte `.param`, and the kernel applied `cvta.shared.u64` to +the host-supplied bytes as if they were a shared address. + +The compiler must reject the signature at compile time instead: + +```bash +cargo oxide build error_kernel_shared_ptr_param +``` + +Expected diagnostic (pinned by `cuda-oxide/scripts/smoketest.sh`): + +```text +holds a pointer into shared memory +``` + +The diagnostic names the field path inside the parameter (`.0` here), which +keeps it distinct from the bare parameter's `is a pointer into shared memory`. +Structs of global (`addrspace(1)`) and constant (`addrspace(4)`) pointers stay +allowed as kernel parameters: the host allocates in those spaces, so it has an +address to supply. A `SharedPtr` belongs on the device side, created inside +the kernel from a `static mut SharedArray` or from `DynamicSharedArray`. diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/src/main.rs b/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/src/main.rs new file mode 100644 index 000000000..74f8b80c7 --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/error_kernel_shared_ptr_param/src/main.rs @@ -0,0 +1,51 @@ +/* + * SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. + * SPDX-License-Identifier: Apache-2.0 + */ + +//! Negative test: a `#[kernel]` parameter struct that holds a `SharedPtr`. +//! +//! A kernel receives its parameters in `.param` space, filled by the host at +//! launch. Shared memory is allocated per block by the device, so the host +//! holds no shared-memory address it could pass. A bare `SharedPtr` +//! parameter is refused for that reason (`error_kernel_shared_param` covers +//! the bare case with `*mut Barrier`), and a struct or array that carries one +//! must be refused the same way. Before the exporter looked inside +//! aggregates, `Args { p: SharedPtr, n: u32 }` compiled and the kernel +//! converted the host-supplied bytes with `cvta.shared.u64` as if they were a +//! shared address. +//! +//! Usage: +//! cargo oxide run error_kernel_shared_ptr_param +//! +//! Expected: the build FAILS with a diagnostic containing (pinned): +//! +//! ```text +//! holds a pointer into shared memory +//! ``` + +use cuda_device::{SharedPtr, kernel}; + +/// Kernel arguments that carry a typed shared pointer next to a plain field. +#[derive(Clone, Copy)] +pub struct Args { + pub p: SharedPtr, + pub n: u32, +} + +/// # Safety +/// +/// Never launched: the device compilation is expected to fail before a host +/// binary that could launch it exists. +#[kernel] +pub unsafe fn shared_ptr_param_kernel(args: Args) { + // BUG UNDER TEST: `args.p` points into shared memory, which the host + // cannot supply at launch. The refusal fires on the entry signature + // alone; the body deliberately does nothing else, so the diagnostic + // under test is the only error this fixture can produce. + let _ = args; +} + +fn main() { + println!("This negative example should fail during device compilation."); +} diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/README.md b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/README.md index 8947f5a80..a022732b0 100644 --- a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/README.md +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/README.md @@ -13,6 +13,58 @@ unsafe fn helper(smem: *mut u8, offset: usize) -> u64 { } ``` +The no-feature build also runs the same probe through `SharedPtr`, a +CTA-shared pointer typed as such (the feature builds fail MIR verification +before any probe runs): + +```rust +let smem = DynamicSharedArray::::shared_ptr(); +shared_ptr_probe(Region { base: smem, offset }); + +#[inline(never)] +unsafe fn shared_ptr_probe(region: Region) -> u64 { + let full = region.base.cast::().byte_add(1024).cast::().as_ptr(); + // ... +} +``` + +`SharedPtr` lowers to an `addrspace(3)` pointer by its type, so the +generic helper's parameter, nested in a struct, is one too, and the call +needs no address-space adaptation. + +Two more shapes run after that probe, in `shared_ptr_shapes_probe`: a struct +with two `SharedPtr` fields passed by value to the non-inlined +`write_pair`, which writes through both, and a `SharedPtr` returned from +the non-inlined `select_slot` and used as a store target. Each of 64 lanes +writes its own slots, then overwrites one slot of its partner lane through the +returned pointer, so the final values prove both stores landed. + +`verify-code-shape.sh` checks the LLVM IR and PTX the build leaves next to +this file: + +| Helper | Required code shape | +| --- | --- | +| `shared_ptr_probe` | first parameter is `ptr addrspace(3)`, and its call site passes one | +| `write_pair` | both `Pair` fields arrive as `ptr addrspace(3)` parameters | +| `select_slot` | return type is `ptr addrspace(3)` | +| `shared_ptr_probe` in PTX | a `.func` that the kernel calls once, with `mbarrier.init.shared` inside its body | +| whole module | no `cvta.to.shared` and no `isspacep` in the PTX | + +The PTX half reads the unoptimized build. `scripts/smoketest.sh` lists this +example in `NO_OPT_SHAPE_EXAMPLES`, so after the optimized build decides the +verdict it rebuilds once with `CUDA_OXIDE_NO_OPT=1` and runs the script on +those artifacts. To run the script by hand, build the same way first: + +```bash +CUDA_OXIDE_NO_OPT=1 cargo oxide build shared_memory_helper +bash crates/rustc-codegen-cuda/examples/shared_memory_helper/verify-code-shape.sh +``` + +Without the surviving `.func`, a module-wide search for `cvta.to.shared` +proves little: an optimized build inlines the helper, and the generic-pointer +control compiles to the same instructions. The call into the `.func` is where +a generic pointer would have needed a conversion. + The launch supplies 2048 dynamic shared bytes. Four independent CTAs each use one thread. Nine launches cover three data offsets and three input values, including wrapping arithmetic. Every CTA checks both a volatile shared-memory @@ -79,6 +131,11 @@ RTX 5090, NVIDIA driver 580.173.02, and CUDA tools 13.3. | Local helper, all three inline policies | MIR verification rejects the call | | Dependency helper, all three inline policies | Same rejection | +With no features, the `SharedPtr` probe runs after the direct control and +reports `PASS shared-ptr: 9 launches, 36 CTA data/barrier checks`, and the +shapes probe reports `PASS shared-ptr-shapes: 3 launches, 384 slot checks`, +on the LLVM NVPTX route and with `--emit-nvvm-ir --arch sm_120`. + Both Compute Sanitizer runs of the direct control reported: ```text diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/src/main.rs b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/src/main.rs index 3abc8dc97..6b25119cb 100644 --- a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/src/main.rs +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/src/main.rs @@ -9,7 +9,8 @@ use cuda_core::simt::LaunchConfig; use cuda_core::{CudaContext, DeviceBuffer}; -use cuda_device::{DynamicSharedArray, cuda_module, kernel, launch_bounds, thread}; +use cuda_device::{DynamicSharedArray, SharedPtr, barrier::Barrier}; +use cuda_device::{cuda_module, kernel, launch_bounds, thread}; #[cfg(all(feature = "local-helper", feature = "cross-crate-helper"))] compile_error!("Select at most one helper feature; no features selects the direct control"); @@ -27,6 +28,82 @@ unsafe fn local_probe(smem: *mut u8, offset: usize) -> u64 { shared_memory_helper_lib::shared_probe_body!(smem, offset) } +/// The shared memory a typed probe works on, as a struct field. +#[derive(Clone, Copy)] +struct Region { + base: SharedPtr, + offset: usize, +} + +/// The probe body over a typed shared pointer, in a generic helper that is not +/// inlined. Unlike the raw `*mut u8` helpers above, it compiles: `SharedPtr` +/// is a CTA-shared pointer by its type, so the helper's parameter is one too, +/// and the barrier is addressed without converting a generic pointer. +/// +/// # Safety +/// As for `shared_probe_body!`: 2048 live shared bytes aligned to 1024, one +/// thread per CTA, and an initialized u64 at `offset`, below byte 1024. +#[inline(never)] +unsafe fn shared_ptr_probe(region: Region) -> u64 { + unsafe { + let bytes = region.base.cast::(); + let full: *mut Barrier = bytes.byte_add(1024).cast::().as_ptr(); + let data = bytes.byte_add(region.offset).cast::().as_ptr(); + cuda_device::barrier::mbarrier_init(full, 1); + let pending = !cuda_device::barrier::mbarrier_try_wait_parity(full, 0); + let _token = cuda_device::barrier::mbarrier_arrive(full); + let mut complete = 0u64; + let mut attempts = 0u32; + while attempts < 1024 { + if cuda_device::barrier::mbarrier_try_wait_parity(full, 0) { + complete = if pending { 1 } else { 2 }; + break; + } + attempts += 1; + } + let value = core::ptr::read_volatile(data); + core::ptr::write_volatile(data, value.wrapping_mul(3) ^ 0x1234_5678); + complete + } +} + +/// Two typed shared pointers in one struct, passed by value. +#[derive(Clone, Copy)] +struct Pair { + a: SharedPtr, + b: SharedPtr, +} + +/// Writes through both fields of a by-value `Pair`. The call boundary splits +/// the struct into two parameters; both must stay CTA-shared pointers. +/// +/// # Safety +/// Both fields must point at live shared memory with at least `index + 1` +/// `f32` slots. +#[inline(never)] +unsafe fn write_pair(pair: Pair, index: usize, value: f32) { + unsafe { + pair.a.add(index).as_ptr().write(value); + pair.b.add(index).as_ptr().write(value + 1000.0); + } +} + +/// Returns one slot of a `Pair` as a typed shared pointer. The return type +/// must stay a CTA-shared pointer so the caller's store needs no conversion. +/// +/// # Safety +/// As for `write_pair`. +#[inline(never)] +unsafe fn select_slot(pair: Pair, index: usize, second: bool) -> SharedPtr { + unsafe { + if second { + pair.b.add(index) + } else { + pair.a.add(index) + } + } +} + #[cuda_module] mod kernels { use super::*; @@ -52,6 +129,54 @@ mod kernels { *output.add(block * 2 + 1) = completed; } } + + /// The same probe through `SharedPtr`, a typed CTA-shared pointer, passed + /// to a non-inlined generic helper inside a struct. + #[kernel] + #[launch_bounds(1)] + pub fn shared_ptr_memory_probe(output: *mut u64, seed: u64, byte_offset: u64) { + let smem = DynamicSharedArray::::shared_ptr(); + let block = thread::blockIdx_x() as usize; + let offset = byte_offset as usize; + unsafe { + let data = smem.byte_add(offset).cast::().as_ptr(); + core::ptr::write_volatile(data, seed.wrapping_add(block as u64)); + let completed = shared_ptr_probe(Region { base: smem, offset }); + *output.add(block * 2) = core::ptr::read_volatile(data); + *output.add(block * 2 + 1) = completed; + } + } + + /// `SharedPtr` shapes a helper boundary must keep: a two-field struct of + /// typed shared pointers passed by value, and a typed shared pointer + /// returned from a helper and used as a store target. + #[kernel] + #[launch_bounds(64)] + pub fn shared_ptr_shapes_probe(output: *mut f32, seed: f32) { + const LANES: usize = 64; + let lane = thread::threadIdx_x() as usize; + let smem = DynamicSharedArray::::shared_ptr(); + // SAFETY: the launch supplies 2048 shared bytes, enough for the two + // 64-slot regions, and every lane writes only its own index and one + // slot of its partner lane, after a barrier. + unsafe { + let pair = Pair { + a: smem, + b: smem.add(LANES), + }; + let value = seed + lane as f32; + write_pair(pair, lane, value); + thread::sync_threads(); + // Each lane overwrites one slot of its partner lane through the + // pointer a helper returns; odd lanes pick the second field. + let partner = LANES - 1 - lane; + let slot = select_slot(pair, partner, lane % 2 == 1); + slot.as_ptr().write(-value); + thread::sync_threads(); + *output.add(lane) = *pair.a.add(lane).as_ptr(); + *output.add(LANES + lane) = *pair.b.add(lane).as_ptr(); + } + } } fn main() -> Result<(), Box> { @@ -79,41 +204,85 @@ fn main() -> Result<(), Box> { block_dim: (1, 1, 1), shared_mem_bytes: 2048, }; - let mut cases = 0; - for seed in [0u64, 17, u64::MAX] { - for offset in [0u64, 64, 248] { - // SAFETY: four independent CTAs, one thread each; output holds two - // u64s per CTA. Data offsets are aligned and disjoint from the - // barrier. Every launch receives a fresh shared-memory allocation. - unsafe { - module.shared_memory_probe( - &stream, - config, - output.cu_deviceptr() as *mut u64, - seed, - offset, - )?; - } - let got = output.to_host_vec(&stream)?; - for block in 0..BLOCKS { - let expected = seed.wrapping_add(block as u64).wrapping_mul(3) ^ 0x1234_5678; - assert_eq!( - got[block * 2], - expected, - "{variant}: data, seed={seed}, offset={offset}, block={block}" - ); - assert_eq!( - got[block * 2 + 1], - 1, - "{variant}: barrier, seed={seed}, offset={offset}, block={block}" - ); + for typed in [false, true] { + let variant = if typed { "shared-ptr" } else { variant }; + let mut cases = 0; + for seed in [0u64, 17, u64::MAX] { + for offset in [0u64, 64, 248] { + let out = output.cu_deviceptr() as *mut u64; + // SAFETY: four independent CTAs, one thread each; output holds + // two u64s per CTA. Data offsets are aligned and disjoint from + // the barrier. Every launch receives a fresh shared-memory + // allocation. + unsafe { + if typed { + module.shared_ptr_memory_probe(&stream, config, out, seed, offset)?; + } else { + module.shared_memory_probe(&stream, config, out, seed, offset)?; + } + } + let got = output.to_host_vec(&stream)?; + for block in 0..BLOCKS { + let expected = seed.wrapping_add(block as u64).wrapping_mul(3) ^ 0x1234_5678; + assert_eq!( + got[block * 2], + expected, + "{variant}: data, seed={seed}, offset={offset}, block={block}" + ); + assert_eq!( + got[block * 2 + 1], + 1, + "{variant}: barrier, seed={seed}, offset={offset}, block={block}" + ); + } + cases += 1; } - cases += 1; + } + println!( + "PASS {variant}: {cases} launches, {} CTA data/barrier checks", + cases * BLOCKS + ); + } + + const LANES: usize = 64; + let shapes_output = DeviceBuffer::::zeroed(&stream, LANES * 2)?; + let shapes_config = LaunchConfig { + grid_dim: (1, 1, 1), + block_dim: (LANES as u32, 1, 1), + shared_mem_bytes: 2048, + }; + let seeds = [0.0f32, 17.0, 1024.0]; + for seed in seeds { + let out = shapes_output.cu_deviceptr() as *mut f32; + // SAFETY: one CTA of 64 lanes; output holds two f32 per lane, and the + // kernel uses 512 of the 2048 shared bytes supplied. + unsafe { module.shared_ptr_shapes_probe(&stream, shapes_config, out, seed)? }; + let got = shapes_output.to_host_vec(&stream)?; + for lane in 0..LANES { + let partner = LANES - 1 - lane; + // Lane `partner` overwrote one of this lane's slots: the second + // field when `partner` is odd, the first field otherwise. + let overwritten = -(seed + partner as f32); + let (want_a, want_b) = if partner % 2 == 1 { + (seed + lane as f32, overwritten) + } else { + (overwritten, seed + lane as f32 + 1000.0) + }; + assert_eq!( + got[lane], want_a, + "shared-ptr-shapes: field a, seed={seed}, lane={lane}" + ); + assert_eq!( + got[LANES + lane], + want_b, + "shared-ptr-shapes: field b, seed={seed}, lane={lane}" + ); } } println!( - "PASS {variant}: {cases} launches, {} CTA data/barrier checks", - cases * BLOCKS + "PASS shared-ptr-shapes: {} launches, {} slot checks", + seeds.len(), + seeds.len() * LANES * 2 ); Ok(()) } diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/verify-code-shape.sh b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/verify-code-shape.sh new file mode 100755 index 000000000..5bc0dc061 --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_memory_helper/verify-code-shape.sh @@ -0,0 +1,103 @@ +#!/usr/bin/env bash +# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. +# SPDX-License-Identifier: Apache-2.0 +# +# `SharedPtr` keeps a CTA-shared pointer typed `addrspace(3)` across a call. +# +# LLVM IR: the non-inlined `shared_ptr_probe` helper takes one as its +# parameter and its call site passes one, the two fields of a by-value `Pair` +# reach `write_pair` as two such parameters, and `select_slot` returns one. +# +# PTX: `shared_ptr_probe` survives as a `.func` that the kernel calls once, +# and that `.func` initializes the barrier with `mbarrier.init.shared` on its +# parameter. No instruction converts a generic address to a shared one +# (`cvta.to.shared`) or asks which window an address is in (`isspacep`). The +# call into the `.func` is where a generic pointer would have needed one; a +# module-wide search alone proves little once the helper is inlined, because +# the generic-pointer control compiles to the same instructions. +# +# The PTX half reads the unoptimized build. scripts/smoketest.sh lists this +# example in NO_OPT_SHAPE_EXAMPLES and rebuilds it with CUDA_OXIDE_NO_OPT=1 +# before running this script; by hand, build the same way first. +set -euo pipefail + +dir="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" +ll="${dir}/shared_memory_helper.ll" +ptx="${dir}/shared_memory_helper.ptx" + +for file in "${ll}" "${ptx}"; do + [[ -f "${file}" ]] || { echo "FAIL: ${file} not found; build the example first" >&2; exit 1; } +done + +# --- LLVM IR ----------------------------------------------------------------- + +if ! grep -Eq '^define .*shared_ptr_probe.*\(ptr addrspace\(3\) ' "${ll}"; then + echo "FAIL: shared_ptr_probe does not take a ptr addrspace(3) parameter" >&2 + grep -E '^define .*shared_ptr_probe' "${ll}" >&2 || true + exit 1 +fi + +if ! grep -Eq 'call i64 @[^(]*shared_ptr_probe[^(]*\(ptr addrspace\(3\) ' "${ll}"; then + echo "FAIL: the call to shared_ptr_probe does not pass a ptr addrspace(3)" >&2 + grep -E 'call i64 @[^(]*shared_ptr_probe' "${ll}" >&2 || true + exit 1 +fi + +if ! grep -Eq '^define .*write_pair.*\(ptr addrspace\(3\) %[A-Za-z0-9_.]+, ptr addrspace\(3\) %[A-Za-z0-9_.]+, ' "${ll}"; then + echo "FAIL: write_pair does not take both Pair fields as ptr addrspace(3) parameters" >&2 + grep -E '^define .*write_pair' "${ll}" >&2 || true + exit 1 +fi + +if ! grep -Eq '^define ptr addrspace\(3\) @[^(]*select_slot' "${ll}"; then + echo "FAIL: select_slot does not return ptr addrspace(3)" >&2 + grep -E '^define .*select_slot' "${ll}" >&2 || true + exit 1 +fi + +# --- PTX --------------------------------------------------------------------- + +# The body of the PTX `.func` whose name contains the symbol: from the header +# line, past the parameter list, to the closing brace. A forward declaration +# ends in `;` before any `{` and is skipped. +func_body() { + awk -v needle="$1" ' + !emit && !candidate && index($0, ".func") && index($0, needle) { candidate = 1 } + candidate && $0 ~ /^[[:space:]]*;[[:space:]]*$/ { candidate = 0; next } + candidate && $0 ~ /^[[:space:]]*\{/ { emit = 1; candidate = 0 } + emit { print } + emit && $0 ~ /^\}/ { exit } + ' "${ptx}" +} + +probe_body="$(func_body shared_ptr_probe)" +if [[ -z "${probe_body}" ]]; then + echo "FAIL: no PTX .func for shared_ptr_probe in ${ptx}" >&2 + echo "hint: this half reads the unoptimized build; rebuild with" >&2 + echo " CUDA_OXIDE_NO_OPT=1 cargo oxide build shared_memory_helper" >&2 + exit 1 +fi + +if ! grep -Eq 'mbarrier\.init\.shared(::cta)?\.b64' <<<"${probe_body}"; then + echo "FAIL: the shared_ptr_probe .func does not initialize the barrier with mbarrier.init.shared" >&2 + exit 1 +fi + +# One call into the helper: the symbol on a `call` line, outside `.func` +# headers and comments. +calls="$(grep -E 'shared_ptr_probe' "${ptx}" | grep -vE '\.func|^[[:space:]]*//' | grep -cE '(^|[[:space:]])call' || true)" +if [[ "${calls}" -ne 1 ]]; then + echo "FAIL: expected one PTX call into shared_ptr_probe, found ${calls}" >&2 + grep -nE 'call.*shared_ptr_probe' "${ptx}" >&2 || true + exit 1 +fi + +for forbidden in 'cvta\.to\.shared' 'isspacep'; do + if grep -Eq "${forbidden}" "${ptx}"; then + echo "FAIL: PTX contains ${forbidden}: a generic address was converted or classified" >&2 + grep -nE "${forbidden}" "${ptx}" >&2 + exit 1 + fi +done + +echo "PASS: SharedPtr stays addrspace(3) across shared_ptr_probe, write_pair and select_slot; the shared_ptr_probe .func is called once and initializes its barrier in shared memory; no cvta.to.shared or isspacep" diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/Cargo.lock b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/Cargo.lock new file mode 100644 index 000000000..0df03ea52 --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/Cargo.lock @@ -0,0 +1,753 @@ +# This file is automatically @generated by Cargo. +# It is not intended for manual editing. +version = 4 + +[[package]] +name = "aho-corasick" +version = "1.1.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ddd31a130427c27518df266943a5308ed92d4b226cc639f5a8f1002816174301" +dependencies = [ + "memchr", +] + +[[package]] +name = "anyhow" +version = "1.0.102" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7f202df86484c868dbad7eaa557ef785d5c66295e41b460ef922eca0723b842c" + +[[package]] +name = "bindgen" +version = "0.69.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "271383c67ccabffb7381723dea0672a673f292304fcb45c01cc648c7a8d58088" +dependencies = [ + "bitflags", + "cexpr", + "clang-sys", + "itertools", + "lazy_static", + "lazycell", + "log", + "prettyplease", + "proc-macro2", + "quote", + "regex", + "rustc-hash", + "shlex", + "syn", + "which", +] + +[[package]] +name = "bitflags" +version = "2.13.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b4388bee8683e3d04af747c73422af53102d2bd24d9eadb6cbc100baef4b43f8" + +[[package]] +name = "block-buffer" +version = "0.10.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3078c7629b62d3f0439517fa394996acacc5cbc91c5a20d8c658e77abd503a71" +dependencies = [ + "generic-array", +] + +[[package]] +name = "cexpr" +version = "0.6.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6fac387a98bb7c37292057cffc56d62ecb629900026402633ae9160df93a8766" +dependencies = [ + "nom", +] + +[[package]] +name = "cfg-if" +version = "1.0.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9330f8b2ff13f34540b44e946ef35111825727b38d33286ef986142615121801" + +[[package]] +name = "clang-sys" +version = "1.8.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0b023947811758c97c59bf9d1c188fd619ad4718dcaa767947df1cadb14f39f4" +dependencies = [ + "glob", + "libc", + "libloading", +] + +[[package]] +name = "cpufeatures" +version = "0.2.17" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "59ed5838eebb26a2bb2e58f6d5b5316989ae9d08bab10e0e6d103e656d1b0280" +dependencies = [ + "libc", +] + +[[package]] +name = "crc32fast" +version = "1.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9481c1c90cbf2ac953f07c8d4a58aa3945c425b7185c9154d67a65e4230da511" +dependencies = [ + "cfg-if", +] + +[[package]] +name = "crunchy" +version = "0.2.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "460fbee9c2c2f33933d720630a6a0bac33ba7053db5344fac858d4b8952d77d5" + +[[package]] +name = "crypto-common" +version = "0.1.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "78c8292055d1c1df0cce5d180393dc8cce0abec0a7102adb6c7b1eef6016d60a" +dependencies = [ + "generic-array", + "typenum", +] + +[[package]] +name = "cuda-artifact-finalizer" +version = "0.2.1" +dependencies = [ + "libnvvm-sys", + "nvjitlink-sys", + "serde", + "sha2", + "thiserror", +] + +[[package]] +name = "cuda-bindings" +version = "0.4.0" +dependencies = [ + "bindgen", + "libloading", + "prettyplease", + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "cuda-core" +version = "0.4.0" +dependencies = [ + "anyhow", + "cuda-bindings", + "cuda-core-derive", + "half", + "oxide-artifacts", +] + +[[package]] +name = "cuda-core-derive" +version = "0.4.0" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "cuda-device" +version = "0.2.1" +dependencies = [ + "cuda-macros", +] + +[[package]] +name = "cuda-host" +version = "0.2.1" +dependencies = [ + "cuda-artifact-finalizer", + "cuda-core", + "cuda-macros", + "half", + "libc", + "oxide-artifacts", + "ptx-parse", + "sha2", + "thiserror", +] + +[[package]] +name = "cuda-macros" +version = "0.2.1" +dependencies = [ + "oxide-tile-export", + "proc-macro2", + "quote", + "reserved-oxide-symbols", + "syn", +] + +[[package]] +name = "cuda-target-spec" +version = "0.1.0" + +[[package]] +name = "digest" +version = "0.10.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9ed9a281f7bc9b7576e61468ba615a66a5c8cfdff42420a70aa82701a3b1e292" +dependencies = [ + "block-buffer", + "crypto-common", +] + +[[package]] +name = "either" +version = "1.16.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "91622ff5e7162018101f2fea40d6ebf4a78bbe5a49736a2020649edf9693679e" + +[[package]] +name = "equivalent" +version = "1.0.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "877a4ace8713b0bcf2a4e7eec82529c029f1d0619886d18145fea96c3ffe5c0f" + +[[package]] +name = "errno" +version = "0.3.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "39cab71617ae0d63f51a36d69f866391735b51691dbda63cf6f96d042b63efeb" +dependencies = [ + "libc", + "windows-sys 0.61.2", +] + +[[package]] +name = "foldhash" +version = "0.1.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d9c4f5dac5e15c24eb999c26181a6ca40b39fe946cbe4c263c7209467bc83af2" + +[[package]] +name = "generic-array" +version = "0.14.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "85649ca51fd72272d7821adaf274ad91c288277713d9c18820d8499a7ff69e9a" +dependencies = [ + "typenum", + "version_check", +] + +[[package]] +name = "glob" +version = "0.3.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0cc23270f6e1808e30a928bdc84dea0b9b4136a8bc82338574f23baf47bbd280" + +[[package]] +name = "half" +version = "2.7.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6ea2d84b969582b4b1864a92dc5d27cd2b77b622a8d79306834f1be5ba20d84b" +dependencies = [ + "cfg-if", + "crunchy", + "zerocopy", +] + +[[package]] +name = "hashbrown" +version = "0.15.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9229cfe53dfd69f0609a49f65461bd93001ea1ef889cd5529dd176593f5338a1" +dependencies = [ + "foldhash", +] + +[[package]] +name = "hashbrown" +version = "0.17.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ed5909b6e89a2db4456e54cd5f673791d7eca6732202bbf2a9cc504fe2f9b84a" + +[[package]] +name = "home" +version = "0.5.12" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cc627f471c528ff0c4a49e1d5e60450c8f6461dd6d10ba9dcd3a61d3dff7728d" +dependencies = [ + "windows-sys 0.61.2", +] + +[[package]] +name = "indexmap" +version = "2.14.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d466e9454f08e4a911e14806c24e16fba1b4c121d1ea474396f396069cf949d9" +dependencies = [ + "equivalent", + "hashbrown 0.17.1", +] + +[[package]] +name = "itertools" +version = "0.12.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ba291022dbbd398a455acf126c1e341954079855bc60dfdda641363bd6922569" +dependencies = [ + "either", +] + +[[package]] +name = "itoa" +version = "1.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682" + +[[package]] +name = "lazy_static" +version = "1.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe" + +[[package]] +name = "lazycell" +version = "1.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55" + +[[package]] +name = "libc" +version = "0.2.186" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "68ab91017fe16c622486840e4c83c9a37afeff978bd239b5293d61ece587de66" + +[[package]] +name = "libloading" +version = "0.8.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d7c4b02199fee7c5d21a5ae7d8cfa79a6ef5bb2fc834d6e9058e89c825efdc55" +dependencies = [ + "cfg-if", + "windows-link", +] + +[[package]] +name = "libnvvm-sys" +version = "0.2.1" +dependencies = [ + "cuda-target-spec", + "libloading", + "thiserror", +] + +[[package]] +name = "linux-raw-sys" +version = "0.4.15" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d26c52dbd32dccf2d10cac7725f8eae5296885fb5703b261f7d0a0739ec807ab" + +[[package]] +name = "log" +version = "0.4.29" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5e5032e24019045c762d3c0f28f5b6b8bbf38563a65908389bf7978758920897" + +[[package]] +name = "memchr" +version = "2.8.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cf8baf1c55e62ffcace7a9f06f4bd9cd3f0c4beb022d3b367256b91b87513d98" + +[[package]] +name = "minimal-lexical" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a" + +[[package]] +name = "nom" +version = "7.1.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a" +dependencies = [ + "memchr", + "minimal-lexical", +] + +[[package]] +name = "nvjitlink-sys" +version = "0.2.1" +dependencies = [ + "libloading", + "thiserror", +] + +[[package]] +name = "object" +version = "0.36.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "62948e14d923ea95ea2c7c86c71013138b66525b86bdc08d2dcc262bdb497b87" +dependencies = [ + "crc32fast", + "hashbrown 0.15.5", + "indexmap", + "memchr", +] + +[[package]] +name = "once_cell" +version = "1.21.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9f7c3e4beb33f85d45ae3e3a1792185706c8e16d043238c593331cc7cd313b50" + +[[package]] +name = "oxide-artifacts" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "74d9a941c0c43d2dceaf25f5c9ca4c9f9cceb28dfa5e4e3cfd14e476f003e7c9" +dependencies = [ + "object", +] + +[[package]] +name = "oxide-tile-export" +version = "0.2.1" +dependencies = [ + "serde", + "serde_json", + "thiserror", +] + +[[package]] +name = "prettyplease" +version = "0.2.37" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "479ca8adacdd7ce8f1fb39ce9ecccbfe93a3f1344b3d0d97f20bc0196208f62b" +dependencies = [ + "proc-macro2", + "syn", +] + +[[package]] +name = "proc-macro2" +version = "1.0.106" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8fd00f0bb2e90d81d1044c2b32617f68fcb9fa3bb7640c23e9c748e53fb30934" +dependencies = [ + "unicode-ident", +] + +[[package]] +name = "ptx-parse" +version = "0.1.0" + +[[package]] +name = "quote" +version = "1.0.45" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "41f2619966050689382d2b44f664f4bc593e129785a36d6ee376ddf37259b924" +dependencies = [ + "proc-macro2", +] + +[[package]] +name = "regex" +version = "1.12.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e10754a14b9137dd7b1e3e5b0493cc9171fdd105e0ab477f51b72e7f3ac0e276" +dependencies = [ + "aho-corasick", + "memchr", + "regex-automata", + "regex-syntax", +] + +[[package]] +name = "regex-automata" +version = "0.4.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6e1dd4122fc1595e8162618945476892eefca7b88c52820e74af6262213cae8f" +dependencies = [ + "aho-corasick", + "memchr", + "regex-syntax", +] + +[[package]] +name = "regex-syntax" +version = "0.8.11" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d6f6ff9a378485b298a5286656da665ba74413d36db0979633275d2e708145d4" + +[[package]] +name = "reserved-oxide-symbols" +version = "0.2.1" + +[[package]] +name = "rustc-hash" +version = "1.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2" + +[[package]] +name = "rustix" +version = "0.38.44" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "fdb5bc1ae2baa591800df16c9ca78619bf65c0488b41b96ccec5d11220d8c154" +dependencies = [ + "bitflags", + "errno", + "libc", + "linux-raw-sys", + "windows-sys 0.59.0", +] + +[[package]] +name = "serde" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9a8e94ea7f378bd32cbbd37198a4a91436180c5bb472411e48b5ec2e2124ae9e" +dependencies = [ + "serde_core", + "serde_derive", +] + +[[package]] +name = "serde_core" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "41d385c7d4ca58e59fc732af25c3983b67ac852c1a25000afe1175de458b67ad" +dependencies = [ + "serde_derive", +] + +[[package]] +name = "serde_derive" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d540f220d3187173da220f885ab66608367b6574e925011a9353e4badda91d79" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "serde_json" +version = "1.0.154" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e7e9cc8b1b85264074fbcc02a88680c4096b1e47df8f739dceb03bf482f04bd6" +dependencies = [ + "itoa", + "memchr", + "serde", + "serde_core", + "zmij", +] + +[[package]] +name = "sha2" +version = "0.10.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a7507d819769d01a365ab707794a4084392c824f54a7a6a7862f8c3d0892b283" +dependencies = [ + "cfg-if", + "cpufeatures", + "digest", +] + +[[package]] +name = "shared_ptr_api" +version = "0.1.0" +dependencies = [ + "cuda-core", + "cuda-device", + "cuda-host", +] + +[[package]] +name = "shlex" +version = "1.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64" + +[[package]] +name = "syn" +version = "2.0.117" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e665b8803e7b1d2a727f4023456bbbbe74da67099c585258af0ad9c5013b9b99" +dependencies = [ + "proc-macro2", + "quote", + "unicode-ident", +] + +[[package]] +name = "thiserror" +version = "2.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4288b5bcbc7920c07a1149a35cf9590a2aa808e0bc1eafaade0b80947865fbc4" +dependencies = [ + "thiserror-impl", +] + +[[package]] +name = "thiserror-impl" +version = "2.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ebc4ee7f67670e9b64d05fa4253e753e016c6c95ff35b89b7941d6b856dec1d5" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "typenum" +version = "1.20.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b6f5e870be6c3b371b77fe0ee0bafb859fa4964b4404c27de1d380043c4dda20" + +[[package]] +name = "unicode-ident" +version = "1.0.24" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e6e4313cd5fcd3dad5cafa179702e2b244f760991f45397d14d4ebf38247da75" + +[[package]] +name = "version_check" +version = "0.9.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0b928f33d975fc6ad9f86c8f283853ad26bdd5b10b7f1542aa2fa15e2289105a" + +[[package]] +name = "which" +version = "4.4.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "87ba24419a2078cd2b0f2ede2691b6c66d8e47836da3b6db8265ebad47afbfc7" +dependencies = [ + "either", + "home", + "once_cell", + "rustix", +] + +[[package]] +name = "windows-link" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f0805222e57f7521d6a62e36fa9163bc891acd422f971defe97d64e70d0a4fe5" + +[[package]] +name = "windows-sys" +version = "0.59.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1e38bc4d79ed67fd075bcc251a1c39b32a1776bbe92e5bef1f0bf1f8c531853b" +dependencies = [ + "windows-targets", +] + +[[package]] +name = "windows-sys" +version = "0.61.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ae137229bcbd6cdf0f7b80a31df61766145077ddf49416a728b02cb3921ff3fc" +dependencies = [ + "windows-link", +] + +[[package]] +name = "windows-targets" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9b724f72796e036ab90c1021d4780d4d3d648aca59e491e6b98e725b84e99973" +dependencies = [ + "windows_aarch64_gnullvm", + "windows_aarch64_msvc", + "windows_i686_gnu", + "windows_i686_gnullvm", + "windows_i686_msvc", + "windows_x86_64_gnu", + "windows_x86_64_gnullvm", + "windows_x86_64_msvc", +] + +[[package]] +name = "windows_aarch64_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "32a4622180e7a0ec044bb555404c800bc9fd9ec262ec147edd5989ccd0c02cd3" + +[[package]] +name = "windows_aarch64_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "09ec2a7bb152e2252b53fa7803150007879548bc709c039df7627cabbd05d469" + +[[package]] +name = "windows_i686_gnu" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8e9b5ad5ab802e97eb8e295ac6720e509ee4c243f69d781394014ebfe8bbfa0b" + +[[package]] +name = "windows_i686_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0eee52d38c090b3caa76c563b86c3a4bd71ef1a819287c19d586d7334ae8ed66" + +[[package]] +name = "windows_i686_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "240948bc05c5e7c6dabba28bf89d89ffce3e303022809e73deaefe4f6ec56c66" + +[[package]] +name = "windows_x86_64_gnu" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "147a5c80aabfbf0c7d901cb5895d1de30ef2907eb21fbbab29ca94c5b08b1a78" + +[[package]] +name = "windows_x86_64_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "24d5b23dc417412679681396f2b49f3de8c1473deb516bd34410872eff51ed0d" + +[[package]] +name = "windows_x86_64_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "589f6da84c646204747d1270a2a5661ea66ed1cced2631d546fdfb155959f9ec" + +[[package]] +name = "zerocopy" +version = "0.8.52" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ce1022995ff5ff5d841ad7d994facc23098cd40152f2c1d11cd607c6f530653f" +dependencies = [ + "zerocopy-derive", +] + +[[package]] +name = "zerocopy-derive" +version = "0.8.52" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1ae7f38b72ec2a254e2b87ef277cf2cd4fb97cbebf944faa6f33354da0867930" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "zmij" +version = "1.0.23" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "29666d0abbfad1e3dc4dcf6144730dd3a3ab225bbbdac83319345b1b44ccfc1b" diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/Cargo.toml b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/Cargo.toml new file mode 100644 index 000000000..f7fd6f703 --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/Cargo.toml @@ -0,0 +1,23 @@ +[package] +name = "shared_ptr_api" +version = "0.1.0" +edition = "2024" +authors = ["NVIDIA Corporation"] +license = "Apache-2.0" +description = "SharedPtr through helper returns, references, closures, arrays, Option, shared storage and transmute" + +# Mark as standalone crate (not part of parent workspace) +[workspace] + +# Build with: +# cargo oxide run shared_ptr_api + +[dependencies] +# Device-side intrinsics (thread indexing, barriers, etc.) +cuda-device = { path = "../../../cuda-device" } + +# Host-side utilities (CudaKernel trait, etc.) +cuda-host = { path = "../../../cuda-host" } + +# Host-side CUDA runtime +cuda-core = { path = "../../../../../cuda-core" } diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/README.md b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/README.md new file mode 100644 index 000000000..52917d7c0 --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/README.md @@ -0,0 +1,74 @@ +# shared_ptr_api + +`SharedPtr` is a raw pointer into the issuing block's shared memory whose +address space (`addrspace(3)`) is part of its type. A generic `*mut T` loses +that fact at the first helper boundary; a `SharedPtr` keeps it through +parameters, return values, struct fields, closures and arrays, so the compiler +can use the shared-memory forms of the instructions that take it. This example +moves one through each of those boundaries and checks every value on the host. + +| Kernel | Boundary | Check | +| :------------------ | :-------------------------------------------------------------- | :-------------------------------------- | +| `static_round_trip` | `SharedArray::shared_ptr`, `from_raw(as_ptr())`, negative `offset` | each thread reads back its own slot | +| `closure_capture` | a closure that captured a `SharedPtr`, called through a generic `#[inline(never)]` helper | `3 * tid` | +| `addr_u32_deltas` | `addr_u32` of `shared_ptr_at(64)`, `add(3)`, `byte_add(7)`, and of `get_raw()` through `cvta_generic_to_shared_u32` | `[64, 12, 1, 7]` | +| `behind_references` | `&Region` (a struct holding the pointer) and `&mut SharedPtr` through non-inlined helpers | three reads summed | +| `returned_pointer` | a non-inlined helper returning `SharedPtr` | odd lanes read the pool, even the array | +| `array_by_value` | `[SharedPtr; 2]` passed by value | `7 + tid` | +| `struct_pointee` | `SharedPtr` and `cast::()` to its first field | `a + b + a` | +| `byte_views` | `byte_add` into a `u8` pool, viewed as `f64` and `u16` | `0.25 * tid + tid + 7` | +| `option_pointer` | `Option>` returned and passed | `Some` stores, `None` does not | +| `stored_in_shared` | `SharedArray, 64>`: pointers kept in shared memory | the mirrored slot's value | +| `transmuted` | `transmute` to `usize` and back in non-inlined helpers | the store lands and the stride is 4 | + +The helpers carry `#[inline(never)]`, so each boundary is a real call. In the +LLVM IR the default build leaves next to this file, `smem_fill`, `smem_load`, +`pick`, `maybe` and the closure's `apply` take `ptr addrspace(3)`, +`store_via_array` takes `[2 x ptr addrspace(3)]`, `pick` and `from_bits` +return `ptr addrspace(3)`, and `SLOTS` is a `[64 x ptr addrspace(3)]` global +in address space 3. The PTX converts no generic address to a shared one +except in `from_bits`: the integer view of a shared pointer is its generic +address, as for unions, so the transmute back is one `cvta.to.shared`. + +Structs that mix a `SharedPtr` with a narrower field, such as +`{ p: SharedPtr, x: u32 }`, stay out of this example. On the libNVVM +route a shared pointer is 4 bytes wide, and such a struct is not yet +supported in arrays or in shared memory; the book's shared memory chapter +lists this limit. + +## Build and run + +```bash +cargo oxide run shared_ptr_api +``` + +The compile-only lane of `scripts/smoketest.sh` also compiles this example +through libNVVM (`NVVM_VERIFY_EXAMPLES`), where the shared pointer is 4 bytes +wide: + +```bash +cargo oxide emit-ltoir shared_ptr_api --arch=sm_120 +``` + +## Expected output + +```text +=== SharedPtr API (64 threads) === + ok static_round_trip + ok closure_capture + ok addr_u32_deltas + ok behind_references + ok returned_pointer + ok array_by_value + ok struct_pointee + ok byte_views + ok option_pointer + ok stored_in_shared + ok transmuted +SUCCESS: shared_ptr_api, 11 kernels, 0 mismatches +``` + +## Hardware requirements + +Any CUDA GPU; the kernels use plain shared memory and `sync_threads()`. The +launch supplies 1024 bytes of dynamic shared memory. diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/src/main.rs b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/src/main.rs new file mode 100644 index 000000000..12ba608b2 --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/shared_ptr_api/src/main.rs @@ -0,0 +1,535 @@ +/* + * SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. + * SPDX-License-Identifier: Apache-2.0 + */ + +//! `SharedPtr` across every boundary a helper call can put it behind. +//! +//! `SharedPtr` is a raw pointer into the issuing block's shared memory +//! whose address space is part of its type. Each kernel here moves one +//! through a boundary where a generic `*mut T` would lose that fact, or where +//! the lowering has to rebuild the value from parts: +//! +//! - a `from_raw` round trip and a negative `offset` on a static array, +//! - a closure that captured one, called through a generic helper, +//! - `addr_u32` deltas against `cvta_generic_to_shared_u32(get_raw())`, +//! - `&Region` and `&mut SharedPtr` through non-inlined helpers, +//! - a non-inlined helper that returns one, +//! - `[SharedPtr; 2]` passed by value, +//! - a struct pointee and a `cast` to its first field, +//! - `f64` and `u16` views of a dynamic byte pool, +//! - `Option>` returned and passed, +//! - `SharedPtr` values stored in shared memory itself, +//! - `transmute` to `usize` and back. +//! +//! The host checks every value and prints one SUCCESS line. Structs that mix +//! a `SharedPtr` with a narrower field stay out of this example: on the +//! libNVVM route a shared pointer is 4 bytes wide, and such a struct is not +//! yet supported in arrays or in shared memory. +//! +//! Usage: +//! cargo oxide run shared_ptr_api + +use cuda_core::simt::LaunchConfig; +use cuda_core::{CudaContext, DeviceBuffer}; +use cuda_device::shared::cvta_generic_to_shared_u32; +use cuda_device::{DynamicSharedArray, SharedArray, SharedPtr, cuda_module, kernel, thread}; + +const THREADS: usize = 64; + +/// A pointee with two fields; a `cast` to `f32` reaches the first. +#[derive(Clone, Copy)] +#[repr(C)] +struct Pair { + a: f32, + b: f32, +} + +/// A typed shared pointer next to an element offset, read through `&Region`. +#[derive(Clone, Copy)] +struct Region { + base: SharedPtr, + offset: usize, +} + +static mut TILE_F32: SharedArray = SharedArray::UNINIT; +static mut TILE_U32: SharedArray = SharedArray::UNINIT; +static mut PAIRS: SharedArray = SharedArray::UNINIT; +static mut SLOTS: SharedArray, THREADS> = SharedArray::UNINIT; + +/// Stores through a typed shared pointer inside a non-inlined generic helper. +/// +/// # Safety +/// `p.add(i)` must be in bounds of live shared memory. +#[inline(never)] +unsafe fn smem_fill(p: SharedPtr, i: usize, v: T) { + unsafe { p.add(i).as_ptr().write(v) } +} + +/// Loads through a typed shared pointer inside a non-inlined generic helper. +/// +/// # Safety +/// `p.add(i)` must be in bounds of initialized shared memory. +#[inline(never)] +unsafe fn smem_load(p: SharedPtr, i: usize) -> T { + unsafe { p.add(i).as_ptr().read() } +} + +/// Returns one of two typed shared pointers. The return type lowers to +/// `ptr addrspace(3)`. +#[inline(never)] +fn pick(a: SharedPtr, b: SharedPtr, second: bool) -> SharedPtr { + if second { b } else { a } +} + +/// Calls a closure that captured a `SharedPtr` by value. +#[inline(never)] +fn apply(f: F, i: usize) { + f(i) +} + +/// Reads through `&Region`: a deref, a field of `Region`, then the pointer. +/// +/// # Safety +/// `r.base.add(r.offset + i)` must be in bounds of initialized shared memory. +#[inline(never)] +unsafe fn read_region(r: &Region, i: usize) -> T { + unsafe { r.base.add(r.offset + i).as_ptr().read() } +} + +/// Advances a `SharedPtr` in place through `&mut SharedPtr`. +/// +/// # Safety +/// The advanced pointer must stay inside the same allocation. +#[inline(never)] +unsafe fn bump(slot: &mut SharedPtr, by: usize) { + unsafe { *slot = slot.add(by) } +} + +/// Stores through one of two `SharedPtr`s passed in an array by value. +/// +/// # Safety +/// `ptrs[which].add(i)` must be in bounds of live shared memory. +#[inline(never)] +unsafe fn store_via_array(ptrs: [SharedPtr; 2], which: usize, i: usize, v: T) { + unsafe { ptrs[which].add(i).as_ptr().write(v) } +} + +/// Stores through an `Option>` and reports whether it did. +/// +/// # Safety +/// A `Some` pointer must be in bounds of live shared memory at `i`. +#[inline(never)] +unsafe fn store_opt(p: Option>, i: usize, v: u32) -> u32 { + match p { + Some(p) => { + unsafe { p.add(i).as_ptr().write(v) }; + 1 + } + None => 0, + } +} + +/// Wraps a `SharedPtr` in `Some` or drops it, in a non-inlined helper. +#[inline(never)] +fn maybe(p: SharedPtr, take: bool) -> Option> { + if take { Some(p) } else { None } +} + +/// A `SharedPtr` as its address bits. +#[inline(never)] +fn to_bits(p: SharedPtr) -> usize { + // SAFETY: `SharedPtr` is `repr(transparent)` over a thin pointer. + unsafe { core::mem::transmute::, usize>(p) } +} + +/// Address bits back to a `SharedPtr`. +/// +/// # Safety +/// `bits` must come from `to_bits` on a live `SharedPtr`. +#[inline(never)] +unsafe fn from_bits(bits: usize) -> SharedPtr { + unsafe { core::mem::transmute::>(bits) } +} + +#[cuda_module] +mod kernels { + use super::*; + + /// A static array through `shared_ptr`, a `from_raw` round trip, and a + /// negative `offset` back to this thread's slot. + #[kernel] + pub fn static_round_trip(out: *mut f32) { + let tid = thread::threadIdx_x() as usize; + // SAFETY: TILE_F32 is a `static mut SharedArray` of this kernel; each + // thread writes and reads its own slot, with a barrier between. + unsafe { + let tile = SharedArray::shared_ptr(&raw mut TILE_F32); + smem_fill(tile, tid, tid as f32 + 0.5); + thread::sync_threads(); + let again = SharedPtr::from_raw(tile.as_ptr()); + let last = again.add(THREADS - 1); + let mine = last.offset(-((THREADS - 1 - tid) as isize)); + *out.add(tid) = smem_load(mine, 0); + } + } + + /// A closure that captured a `SharedPtr`, called through a generic + /// non-inlined helper. + #[kernel] + pub fn closure_capture(out: *mut u32) { + let tid = thread::threadIdx_x() as usize; + // SAFETY: TILE_U32 is a `static mut SharedArray` of this kernel. + let tile = unsafe { SharedArray::shared_ptr(&raw mut TILE_U32) }; + // SAFETY: each thread writes its own slot. + let store = move |i: usize| unsafe { tile.add(i).as_ptr().write(i as u32 * 3) }; + apply(store, tid); + thread::sync_threads(); + // SAFETY: the slot was written before the barrier. + unsafe { *out.add(tid) = tile.add(tid).as_ptr().read() }; + } + + /// `addr_u32` deltas within the dynamic pool, and equality with + /// `cvta_generic_to_shared_u32` on the pool's raw pointer. + #[kernel] + pub fn addr_u32_deltas(out: *mut u32) { + let tid = thread::threadIdx_x() as usize; + let base = DynamicSharedArray::::shared_ptr(); + let at64 = DynamicSharedArray::::shared_ptr_at(64); + // SAFETY: `get_raw` points into the dynamic pool; thread 0 writes the + // four output slots. + unsafe { + let raw = cvta_generic_to_shared_u32(DynamicSharedArray::::get_raw()); + if tid == 0 { + *out.add(0) = at64.addr_u32() - base.addr_u32(); + *out.add(1) = base.add(3).addr_u32() - base.addr_u32(); + *out.add(2) = (base.addr_u32() == raw) as u32; + *out.add(3) = base.cast::().byte_add(7).addr_u32() - base.addr_u32(); + } + } + } + + /// `SharedPtr` behind references: `&Region` and `&mut SharedPtr`. + #[kernel] + pub fn behind_references(out: *mut u32) { + let tid = thread::threadIdx_x() as usize; + let pool = DynamicSharedArray::::shared_ptr(); + // SAFETY: the launch supplies 1024 dynamic shared bytes, enough for + // the 64 `u32` slots; every slot is written before the barrier. + unsafe { + pool.add(tid).as_ptr().write(100 + tid as u32); + thread::sync_threads(); + let region = Region { + base: pool, + offset: 0, + }; + let a = read_region(®ion, tid); + let mut cursor = pool; + bump(&mut cursor, tid); + let b = cursor.as_ptr().read(); + let shifted = Region { + base: pool, + offset: THREADS / 2, + }; + let c = read_region(&shifted, tid % (THREADS / 2)); + *out.add(tid) = a + b + c; + } + } + + /// A non-inlined helper that returns a `SharedPtr` picks between a static + /// array and the dynamic pool. + #[kernel] + pub fn returned_pointer(out: *mut f32) { + let tid = thread::threadIdx_x() as usize; + let dynp = DynamicSharedArray::::shared_ptr(); + // SAFETY: TILE_F32 is a `static mut SharedArray` of this kernel and + // the pool holds 64 `f32`; each thread writes its own slot in both. + unsafe { + let tile = SharedArray::shared_ptr(&raw mut TILE_F32); + smem_fill(tile, tid, 1000.0 + tid as f32); + smem_fill(dynp, tid, 2000.0 + tid as f32); + thread::sync_threads(); + let chosen = pick(tile, dynp, tid % 2 == 1); + *out.add(tid) = smem_load(chosen, tid); + } + } + + /// `[SharedPtr; 2]` passed by value to a non-inlined helper. + #[kernel] + pub fn array_by_value(out: *mut u32) { + let tid = thread::threadIdx_x() as usize; + let pool = DynamicSharedArray::::shared_ptr(); + // SAFETY: TILE_U32 is a `static mut SharedArray` of this kernel and + // the pool holds 64 `u32`; each thread writes its own slot in one. + unsafe { + let tile = SharedArray::shared_ptr(&raw mut TILE_U32); + store_via_array([tile, pool], tid % 2, tid, 7 + tid as u32); + thread::sync_threads(); + let v = if tid.is_multiple_of(2) { + tile.add(tid).as_ptr().read() + } else { + pool.add(tid).as_ptr().read() + }; + *out.add(tid) = v; + } + } + + /// A struct pointee, and a `cast` to its first field. + #[kernel] + pub fn struct_pointee(out: *mut f32) { + let tid = thread::threadIdx_x() as usize; + // SAFETY: PAIRS is a `static mut SharedArray` of this kernel; each + // thread writes and reads its own element. + unsafe { + let pairs = SharedArray::shared_ptr(&raw mut PAIRS); + smem_fill( + pairs, + tid, + Pair { + a: tid as f32, + b: 2.0 * tid as f32, + }, + ); + thread::sync_threads(); + let p = smem_load(pairs, tid); + let first: f32 = smem_load(pairs.add(tid).cast::(), 0); + *out.add(tid) = p.a + p.b + first; + } + } + + /// `byte_add` into a dynamic byte pool, viewed as `f64` and as `u16`. + #[kernel] + pub fn byte_views(out: *mut f64) { + let tid = thread::threadIdx_x() as usize; + let pool = DynamicSharedArray::::shared_ptr(); + // SAFETY: the pool holds 1024 bytes; the `f64` view uses the first + // 512 and the `u16` view the next 128, each thread its own slot. + unsafe { + let d = pool.byte_add(8 * tid).cast::(); + smem_fill(d, 0, tid as f64 * 0.25); + let h = pool.byte_add(8 * THREADS + 2 * tid).cast::(); + smem_fill(h, 0, tid as u16 + 7); + thread::sync_threads(); + let dv: f64 = smem_load(pool.cast::(), tid); + let hv: u16 = smem_load(pool.byte_add(8 * THREADS).cast::(), tid); + *out.add(tid) = dv + hv as f64; + } + } + + /// `Option>` returned from and passed into non-inlined + /// helpers. + #[kernel] + pub fn option_pointer(out: *mut u32) { + let tid = thread::threadIdx_x() as usize; + // SAFETY: TILE_U32 is a `static mut SharedArray` of this kernel; each + // thread writes and reads its own slot. + unsafe { + let tile = SharedArray::shared_ptr(&raw mut TILE_U32); + let took = store_opt(maybe(tile, true), tid, 5 + tid as u32); + let skipped = store_opt(maybe(tile, false), tid, 0); + thread::sync_threads(); + *out.add(tid) = tile.add(tid).as_ptr().read() + 10 * took + 100 * skipped; + } + } + + /// `SharedPtr` values stored in shared memory: a `SharedArray` of them. + #[kernel] + pub fn stored_in_shared(out: *mut u32) { + let tid = thread::threadIdx_x() as usize; + // SAFETY: TILE_U32 and SLOTS are `static mut SharedArray`s of this + // kernel; each thread writes its own slot in both and reads one slot + // after the barrier. + unsafe { + let tile = SharedArray::shared_ptr(&raw mut TILE_U32); + let slots = SharedArray::shared_ptr(&raw mut SLOTS); + tile.add(tid).as_ptr().write(tid as u32 * 11); + slots.add(tid).as_ptr().write(tile.add(THREADS - 1 - tid)); + thread::sync_threads(); + let p: SharedPtr = slots.add(tid).as_ptr().read(); + *out.add(tid) = p.as_ptr().read(); + } + } + + /// `transmute` to `usize` and back, in non-inlined helpers. The high word + /// records whether the bits of this thread's slot sit `4 * tid` bytes + /// past the bits of the array base. + #[kernel] + pub fn transmuted(out: *mut u64) { + let tid = thread::threadIdx_x() as usize; + // SAFETY: TILE_U32 is a `static mut SharedArray` of this kernel; each + // thread writes and reads its own slot. + unsafe { + let tile = SharedArray::shared_ptr(&raw mut TILE_U32); + let bits = to_bits(tile.add(tid)); + let back = from_bits(bits); + back.as_ptr().write(tid as u32 * 13); + thread::sync_threads(); + let v = tile.add(tid).as_ptr().read(); + let strided = bits == to_bits(tile) + 4 * tid; + *out.add(tid) = v as u64 | ((strided as u64) << 32); + } + } +} + +/// Counts the mismatches of one kernel and prints its line. +fn check(name: &str, bad: usize, mismatches: &mut usize, kernels: &mut usize) { + if bad == 0 { + println!(" ok {name}"); + } else { + println!(" BAD {name}: {bad} mismatches"); + } + *mismatches += bad; + *kernels += 1; +} + +fn main() -> Result<(), Box> { + let ctx = CudaContext::new(0)?; + let stream = ctx.default_stream(); + let module = kernels::load(&ctx)?; + let config = LaunchConfig { + grid_dim: (1, 1, 1), + block_dim: (THREADS as u32, 1, 1), + shared_mem_bytes: 1024, + }; + let mut mismatches = 0usize; + let mut kernels = 0usize; + println!("=== SharedPtr API ({THREADS} threads) ==="); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + // SAFETY: one block of THREADS threads; every output buffer holds one + // slot per thread (four slots for addr_u32_deltas, written by thread 0). + unsafe { module.static_round_trip(&stream, config, out.cu_deviceptr() as *mut f32)? }; + let got = out.to_host_vec(&stream)?; + check( + "static_round_trip", + (0..THREADS).filter(|&i| got[i] != i as f32 + 0.5).count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.closure_capture(&stream, config, out.cu_deviceptr() as *mut u32)? }; + let got = out.to_host_vec(&stream)?; + check( + "closure_capture", + (0..THREADS).filter(|&i| got[i] != i as u32 * 3).count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, 4)?; + unsafe { module.addr_u32_deltas(&stream, config, out.cu_deviceptr() as *mut u32)? }; + let got = out.to_host_vec(&stream)?; + let want = [64u32, 12, 1, 7]; + check( + "addr_u32_deltas", + (0..4).filter(|&i| got[i] != want[i]).count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.behind_references(&stream, config, out.cu_deviceptr() as *mut u32)? }; + let got = out.to_host_vec(&stream)?; + check( + "behind_references", + (0..THREADS) + .filter(|&i| { + let t = i as u32; + let half = (THREADS / 2) as u32; + got[i] != (100 + t) + (100 + t) + (100 + half + t % half) + }) + .count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.returned_pointer(&stream, config, out.cu_deviceptr() as *mut f32)? }; + let got = out.to_host_vec(&stream)?; + check( + "returned_pointer", + (0..THREADS) + .filter(|&i| { + let base = if i % 2 == 1 { 2000.0 } else { 1000.0 }; + got[i] != base + i as f32 + }) + .count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.array_by_value(&stream, config, out.cu_deviceptr() as *mut u32)? }; + let got = out.to_host_vec(&stream)?; + check( + "array_by_value", + (0..THREADS).filter(|&i| got[i] != 7 + i as u32).count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.struct_pointee(&stream, config, out.cu_deviceptr() as *mut f32)? }; + let got = out.to_host_vec(&stream)?; + check( + "struct_pointee", + (0..THREADS).filter(|&i| got[i] != 4.0 * i as f32).count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.byte_views(&stream, config, out.cu_deviceptr() as *mut f64)? }; + let got = out.to_host_vec(&stream)?; + check( + "byte_views", + (0..THREADS) + .filter(|&i| got[i] != i as f64 * 0.25 + (i as f64 + 7.0)) + .count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.option_pointer(&stream, config, out.cu_deviceptr() as *mut u32)? }; + let got = out.to_host_vec(&stream)?; + check( + "option_pointer", + (0..THREADS).filter(|&i| got[i] != 15 + i as u32).count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.stored_in_shared(&stream, config, out.cu_deviceptr() as *mut u32)? }; + let got = out.to_host_vec(&stream)?; + check( + "stored_in_shared", + (0..THREADS) + .filter(|&i| got[i] != (THREADS - 1 - i) as u32 * 11) + .count(), + &mut mismatches, + &mut kernels, + ); + + let out = DeviceBuffer::::zeroed(&stream, THREADS)?; + unsafe { module.transmuted(&stream, config, out.cu_deviceptr() as *mut u64)? }; + let got = out.to_host_vec(&stream)?; + check( + "transmuted", + (0..THREADS) + .filter(|&i| got[i] != (i as u64 * 13) | (1u64 << 32)) + .count(), + &mut mismatches, + &mut kernels, + ); + + if mismatches == 0 { + println!("SUCCESS: shared_ptr_api, {kernels} kernels, 0 mismatches"); + Ok(()) + } else { + println!("FAIL: shared_ptr_api, {mismatches} mismatches across {kernels} kernels"); + std::process::exit(1) + } +} diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/README.md b/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/README.md index b89b9503b..0f8b32dba 100644 --- a/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/README.md +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/README.md @@ -10,9 +10,17 @@ For TMA multicast (also sm_90+), see the [`tma_multicast`](../tma_multicast/) ex ## What This Example Does 1. **tma_copy_2d_test**: Async 2D tile copy from global to shared memory -2. **tma_pipeline_test**: TMA with mbarrier for completion tracking +2. **tma_copy_2d_cta_test**: The same copy with `cp_async_bulk_tensor_2d_g2s_cta`, + whose destination is a `SharedPtr`, the issuing CTA's own shared memory by type +3. **tma_pipeline_test**: TMA with mbarrier for completion tracking -Both kernels use `#[grid_constant] tensor_map: &TmaDescriptor`. The generated +`cp_async_bulk_tensor_2d_g2s` addresses cluster shared memory, so its +destination may belong to another CTA of the cluster. The `_cta` form takes a +`SharedPtr` destination, here `SharedArray::shared_ptr(&raw mut TILE)`, and +emits `cp.async.bulk.tensor.2d.shared::cta.global` without converting any +address; `verify-code-shape.sh` checks both forms in the generated PTX. + +All kernels use `#[grid_constant] tensor_map: &TmaDescriptor`. The generated host launcher takes the descriptor by value and places its 128 bytes directly in the kernel's parameter storage. Every thread can address that same read-only value; no separate device descriptor allocation or upload is needed. The tensor diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/src/main.rs b/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/src/main.rs index 04ba8415d..aa21c0644 100644 --- a/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/src/main.rs +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/src/main.rs @@ -9,6 +9,9 @@ //! //! Demonstrates TMA (Tensor Memory Accelerator) usage: //! - cp_async_bulk_tensor_2d_g2s: Async 2D tensor copy global → shared +//! - cp_async_bulk_tensor_2d_g2s_cta: the same copy into the issuing CTA's +//! own shared memory, whose destination is a `SharedPtr`, a CTA-shared +//! pointer by type //! - mbarrier: Barrier-based completion tracking //! - `#[grid_constant]`: the descriptor is passed by value in the launch packet, //! without allocating or uploading a separate descriptor buffer. @@ -34,7 +37,9 @@ use cuda_device::barrier::{ Barrier, fence_proxy_async_shared_cta, mbarrier_arrive, mbarrier_arrive_expect_tx, mbarrier_init, mbarrier_try_wait, }; -use cuda_device::tma::{TmaDescriptor, cp_async_bulk_tensor_2d_g2s}; +use cuda_device::tma::{ + TmaDescriptor, cp_async_bulk_tensor_2d_g2s, cp_async_bulk_tensor_2d_g2s_cta, +}; use cuda_device::{DisjointSlice, SharedArray, kernel, thread}; use cuda_host::cuda_module; use std::mem::MaybeUninit; @@ -125,6 +130,65 @@ mod kernels { } } + /// `tma_copy_2d_test` with the CTA-local copy: the tile and the barrier + /// are this CTA's own shared statics, so the `.shared::cta` form applies. + #[kernel] + pub fn tma_copy_2d_cta_test( + #[grid_constant] tensor_map: &TmaDescriptor, + mut out: DisjointSlice, + tile_x: i32, + tile_y: i32, + ) { + const TILE_SIZE: usize = 64 * 64; + const TILE_BYTES: u32 = (TILE_SIZE * 4) as u32; + static mut TILE: SharedArray = SharedArray::UNINIT; + static mut BAR: Barrier = Barrier::UNINIT; + + let tid = thread::threadIdx_x(); + let block_size = thread::blockDim_x(); + let gid = thread::index_1d(); + + if tid == 0 { + unsafe { + mbarrier_init(&raw mut BAR, block_size); + fence_proxy_async_shared_cta(); + } + } + thread::sync_threads(); + + if tid == 0 { + // SAFETY: TILE and BAR outlive the copy, which completes on BAR + // before any thread reads TILE. + unsafe { + cp_async_bulk_tensor_2d_g2s_cta( + SharedArray::shared_ptr(&raw mut TILE).cast::(), + tensor_map, + tile_x, + tile_y, + &raw mut BAR, + ); + } + } + + let token = unsafe { + if tid == 0 { + mbarrier_arrive_expect_tx(&raw const BAR, 1, TILE_BYTES) + } else { + mbarrier_arrive(&raw const BAR) + } + }; + unsafe { while !mbarrier_try_wait(&raw const BAR, token) {} } + thread::sync_threads(); + + let idx = gid.get(); + if idx < TILE_SIZE { + let val = unsafe { TILE[idx] }; + if let Some(out_elem) = out.get_mut(gid) { + *out_elem = val; + } + } + } + /// Simple TMA pipeline test - ALL threads participate in barrier. #[kernel] pub fn tma_pipeline_test( @@ -227,7 +291,8 @@ fn main() -> Result<(), Box> { println!("✓ Module loaded successfully\n"); // Run tests - run_tma_copy_test(&stream, &module)?; + run_tma_copy_test(&stream, &module, false)?; + run_tma_copy_test(&stream, &module, true)?; run_tma_pipeline_test(&stream, &module)?; println!("\n=== TMA Copy Test Complete ==="); @@ -265,8 +330,14 @@ fn verify_ptx_only(ctx: &Arc) -> Result<(), Box, module: &kernels::LoadedModule, + cta: bool, ) -> Result<(), Box> { - println!("--- Test 1: TMA Copy (tma_copy_2d_test) ---\n"); + let kernel = if cta { + "tma_copy_2d_cta_test" + } else { + "tma_copy_2d_test" + }; + println!("--- Test: TMA Copy ({kernel}) ---\n"); println!( "1. Allocating {}x{} tensor ({} floats, {} KB)", @@ -294,7 +365,7 @@ fn run_tma_copy_test( TILE_HEIGHT, )?; - println!("3. Launching tma_copy_2d_test kernel..."); + println!("3. Launching {kernel} kernel..."); let tile_x: i32 = 0; let tile_y: i32 = 0; @@ -309,14 +380,25 @@ fn run_tma_copy_test( // SAFETY: launch shape/resources match the kernel; buffers cover its accesses. unsafe { - module.tma_copy_2d_test( - (stream).as_ref(), - cfg, - tensor_map, - &mut dev_output, - tile_x, - tile_y, - ) + if cta { + module.tma_copy_2d_cta_test( + (stream).as_ref(), + cfg, + tensor_map, + &mut dev_output, + tile_x, + tile_y, + ) + } else { + module.tma_copy_2d_test( + (stream).as_ref(), + cfg, + tensor_map, + &mut dev_output, + tile_x, + tile_y, + ) + } }?; stream.synchronize()?; @@ -359,7 +441,7 @@ fn run_tma_pipeline_test( stream: &Arc, module: &kernels::LoadedModule, ) -> Result<(), Box> { - println!("\n--- Test 2: TMA Pipeline (tma_pipeline_test) ---\n"); + println!("\n--- Test: TMA Pipeline (tma_pipeline_test) ---\n"); const PIPELINE_TILE_WIDTH: u32 = 32; const PIPELINE_TILE_HEIGHT: u32 = 32; diff --git a/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/verify-code-shape.sh b/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/verify-code-shape.sh new file mode 100755 index 000000000..818506f8f --- /dev/null +++ b/cuda-oxide/crates/rustc-codegen-cuda/examples/tma_copy/verify-code-shape.sh @@ -0,0 +1,75 @@ +#!/usr/bin/env bash +# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. +# SPDX-License-Identifier: Apache-2.0 + +# cp_async_bulk_tensor_2d_g2s_cta must copy into the issuing CTA's shared +# memory: its kernel emits the `.shared::cta` destination form and, since its +# `SharedPtr` destination is shared by type, converts no address to the shared +# or cluster window. The cluster API must keep its +# `.shared::cluster` form, since its destination may belong to another CTA. + +set -euo pipefail + +root="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" +ll="${root}/tma_copy.ll" +ptx="${root}/tma_copy.ptx" + +test -s "${ll}" +test -s "${ptx}" + +# Print the body of one PTX entry. Waits for the header's `{` so a forward +# declaration (which ends in `;`) is skipped. +entry_body() { + local symbol="$1" + awk -v marker="${symbol}(" ' + !emit && !candidate && index($0, marker) && index($0, ".entry") { + candidate = 1 + } + candidate && $0 ~ /^[[:space:]]*;[[:space:]]*$/ { + candidate = 0 + next + } + candidate && $0 ~ /^[[:space:]]*\{[[:space:]]*$/ { + emit = 1 + candidate = 0 + } + emit { print } + emit && index($0, "End function") != 0 { exit } + ' "${ptx}" +} + +require_entry_shape() { + local symbol="$1" + local description="$2" + local pattern="$3" + if ! entry_body "${symbol}" | grep -E "${pattern}" >/dev/null; then + echo "error: missing ${description} in ${ptx}:${symbol}" >&2 + exit 1 + fi +} + +forbid_entry_shape() { + local symbol="$1" + local description="$2" + local pattern="$3" + if entry_body "${symbol}" | grep -E "${pattern}" >/dev/null; then + echo "error: unexpected ${description} in ${ptx}:${symbol}" >&2 + exit 1 + fi +} + +if ! grep -E 'call void @llvm\.nvvm\.cp\.async\.bulk\.tensor\.g2s\.cta\.tile\.2d\(ptr addrspace\(3\)' \ + "${ll}" >/dev/null; then + echo "error: missing typed CTA-local G2S call in ${ll}" >&2 + exit 1 +fi + +require_entry_shape tma_copy_2d_cta_test "CTA-local TMA load" \ + 'cp\.async\.bulk\.tensor\.2d\.shared::cta\.global\.tile\.mbarrier::complete_tx::bytes' +forbid_entry_shape tma_copy_2d_cta_test "cluster-window destination" 'shared::cluster' +forbid_entry_shape tma_copy_2d_cta_test "generic-to-shared address conversion" 'cvta\.to\.shared' + +require_entry_shape tma_copy_2d_test "cluster TMA load" \ + 'cp\.async\.bulk\.tensor\.2d\.shared::cluster\.global\.tile\.mbarrier::complete_tx::bytes' + +echo "tma_copy code shape: PASS" diff --git a/cuda-oxide/crates/rustc-codegen-cuda/src/generated_intrinsics.rs b/cuda-oxide/crates/rustc-codegen-cuda/src/generated_intrinsics.rs index 9ee8fb794..311200a1a 100644 --- a/cuda-oxide/crates/rustc-codegen-cuda/src/generated_intrinsics.rs +++ b/cuda-oxide/crates/rustc-codegen-cuda/src/generated_intrinsics.rs @@ -4,7 +4,7 @@ */ // @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +// Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. // LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. //! Generated collector predicates for intrinsic placeholder functions. @@ -99,16 +99,21 @@ pub(crate) const GENERATED_INTRINSIC_CANONICAL_PATHS: &[&str] = &[ "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0893", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0922", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0328", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0335", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0329", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0330", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0331", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0336", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0332", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0333", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0334", + "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0339", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0928", "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0929", @@ -1135,16 +1140,21 @@ pub(crate) const GENERATED_INTRINSIC_PUBLIC_PATHS: &[&str] = &[ "cuda_intrinsics::tma::cp_async_bulk_prefetch_tensor_gather4_2d_l2", "cuda_intrinsics::tma::cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint", "cuda_intrinsics::tma::cp_async_bulk_tensor_1d_g2s", + "cuda_intrinsics::tma::cp_async_bulk_tensor_1d_g2s_cta", "cuda_intrinsics::tma::cp_async_bulk_tensor_1d_s2g", "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s", + "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_cta", "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_multicast", "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_multicast_cg2", "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_s2g", "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s", + "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s_cta", "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_s2g", "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s", + "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s_cta", "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_s2g", "cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s", + "cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s_cta", "cuda_intrinsics::tma::cp_async_bulk_tensor_5d_s2g", "cuda_intrinsics::tma::cp_async_bulk_tensor_reduce_add_im2col_3d", "cuda_intrinsics::tma::cp_async_bulk_tensor_reduce_add_im2col_4d", @@ -2177,16 +2187,21 @@ pub(crate) fn is_generated_intrinsic_canonical_path(path: &str) -> bool { | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0893" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0922" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0328" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0335" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0329" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0330" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0331" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0336" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0332" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0333" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0334" + | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0339" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0928" | "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0929" @@ -3220,16 +3235,21 @@ pub(crate) fn is_generated_intrinsic_compatibility_path(path: &str) -> bool { | "cuda_device::tma::cp_async_bulk_prefetch_tensor_gather4_2d_l2" | "cuda_device::tma::cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint" | "cuda_device::tma::cp_async_bulk_tensor_1d_g2s" + | "cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta" | "cuda_device::tma::cp_async_bulk_tensor_1d_s2g" | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s" + | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta" | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_multicast" | "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_multicast_cg2" | "cuda_device::tma::cp_async_bulk_tensor_2d_s2g" | "cuda_device::tma::cp_async_bulk_tensor_3d_g2s" + | "cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta" | "cuda_device::tma::cp_async_bulk_tensor_3d_s2g" | "cuda_device::tma::cp_async_bulk_tensor_4d_g2s" + | "cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta" | "cuda_device::tma::cp_async_bulk_tensor_4d_s2g" | "cuda_device::tma::cp_async_bulk_tensor_5d_g2s" + | "cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta" | "cuda_device::tma::cp_async_bulk_tensor_5d_s2g" | "cuda_device::tma::cp_async_bulk_tensor_reduce_add_im2col_3d" | "cuda_device::tma::cp_async_bulk_tensor_reduce_add_im2col_4d" diff --git a/cuda-oxide/cuda-oxide-book/advanced/images/tma-cta-local.svg b/cuda-oxide/cuda-oxide-book/advanced/images/tma-cta-local.svg new file mode 100644 index 000000000..acbd534d5 --- /dev/null +++ b/cuda-oxide/cuda-oxide-book/advanced/images/tma-cta-local.svg @@ -0,0 +1,139 @@ + + + + + +CTA-local TMA loads: a shared pointer that is shared by type, and the .shared::cta copy form +The cluster form converts a generic *mut u8 to the cluster window; on sm_120a ptxas adds a runtime CTA check whose call drops setmaxnreg. The CTA form takes a SharedPtr<u8>. +TODAY (cp_async_bulk_tensor_2d_g2s, generic *mut u8) +THE TYPE (SharedPtr<T>, from #1341) +NEW (cp_async_bulk_tensor_2d_g2s_cta, SharedPtr<u8>) + +tma_copy_2d_test: the cluster form +static mut TILE: SharedArray<f32, TILE_SIZE, 128> + = SharedArray::UNINIT; +static mut BAR: Barrier = Barrier::UNINIT; + +// dst: a generic *mut u8, any CTA of the cluster +cp_async_bulk_tensor_2d_g2s( + &raw mut TILE as *mut u8, + tensor_map, tile_x, tile_y, + &raw mut BAR, +); +Unchanged by this PR: 39 in-tree callers keep this form and +its cluster contract; the PTX is byte for byte the same. + +tma_copy.ll: the generic pointer is converted +%v21 = addrspacecast ptr addrspace(3) @TILE to ptr +%v22 = addrspacecast ptr %v21 to ptr addrspace(7) +call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d( + ptr addrspace(7) %v22, ptr addrspace(3) @BAR, ptr %v7, + i32 %v9, i32 %v10, i16 0, i64 0, i1 0, i1 0, i32 0) +The destination arrives as a generic pointer (the user's cast), +so the lowering can only name the cluster window, addrspace(7). +On sm_100a the cluster form is native. On sm_120a ptxas cannot +prove a non-multicast .shared::cluster destination local, so +it inserts a runtime CTA check with a syscall fallback, and +that call makes it ignore setmaxnreg (C7506): the #1238 kernel +ran with 24 registers instead of 168. + +SharedPtr<T>: addrspace(3) by type +let t = unsafe { SharedArray::shared_ptr(&raw mut TILE) }; +let d = DynamicSharedArray::<u8, 128>::shared_ptr_at(1024); +let p = unsafe { SharedPtr::from_raw(generic) }; // asserts +t.add(i) t.byte_add(b) t.cast::<u8>() // keep the type +t.as_ptr() // the generic *mut T for loads and stores +t.addr_u32() // the 32-bit .shared::cta address +repr(transparent) over *mut SharedSlot<T>; the importer types +that pointee as shared, next to SharedArray and Barrier, and +the newtype lowers as one scalar, like ThreadIndex. + +shared_memory_helper: across helper boundaries +struct Region<T> { base: SharedPtr<T>, offset: usize } +struct Pair { a: SharedPtr<f32>, b: SharedPtr<f32> } +#[inline(never)] +unsafe fn shared_ptr_probe<T>(region: Region<T>) -> u64 +#[inline(never)] +unsafe fn write_pair(pair: Pair, index: usize, value: f32) +#[inline(never)] +unsafe fn select_slot(pair: Pair, index: usize, second: bool) + -> SharedPtr<f32> + +the .ll: what each boundary keeps +define i64 @..shared_ptr_probe..( + ptr addrspace(3) %v0, i64 %v1) +define void @..write_pair( + ptr addrspace(3) %v0, ptr addrspace(3) %v1, i64, float) +define ptr addrspace(3) @..select_slot( + ptr addrspace(3) %v0, ptr addrspace(3) %v1, i64, i1) +Both routes; 0 cvta.to.shared in the PTX. A SharedPtr kernel +parameter, bare or in a struct or array, is a compile error. + +tma_copy_2d_cta_test: the CTA form (new) +// dst: SharedPtr<u8>, this CTA's shared memory by type +cp_async_bulk_tensor_2d_g2s_cta( + SharedArray::shared_ptr(&raw mut TILE).cast::<u8>(), + tensor_map, tile_x, tile_y, + &raw mut BAR, +); +cp_async_bulk_tensor_{1..5}d_g2s_cta, catalog records +i1034 to i1038, PTX 8.6, sm_90 and newer; tile mode only. + +LLVM NVPTX route: typed intrinsic (LLVM 22+) +call void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d( + ptr addrspace(3) @TILE, ptr addrspace(3) @BAR, ptr %v7, + i32 %v9, i32 %v10, i64 0, i1 0) +No addrspacecast: the destination is already addrspace(3). + +libNVVM route: inline PTX, u32 addresses +asm sideeffect + "cp.async.bulk.tensor.2d.shared::cta.global.tile + .mbarrier::complete_tx::bytes [$0], [$2, {$3, $4}], [$1];", + "r,r,l,r,r,~{memory}" + +refused at compile time +a SharedPtr kernel parameter, bare or inside a struct or array +a barrier passed straight from map_shared_rank_mut (a located +error that names the cluster form to use instead) +the caller's contract: from_raw on a cluster pointer; the raw +cuda_intrinsics layer keeps *mut u8 and converts with +cvta.to.shared, and its safety text says CTA-local + + +THE PTX AND WHAT PTXAS MAKES OF IT ON sm_120a (tma_copy.ptx built for sm_90a; mnemonic column aligned; the copy line wrapped before .mbarrier; register counts from ptxas 13.4) + +cluster form: tma_copy_2d_test (unchanged) +mov.b64 %rd15, TILE; +cvta.shared.u64 %rd16, %rd15; +cvta.to.shared::cluster.u64 %rd17, %rd16; +cp.async.bulk.tensor.2d.shared::cluster.global.tile + .mbarrier::complete_tx::bytes [%rd17], [%rd2, {%r3, %r4}], [BAR]; +// ptxas sm_120a: 24 registers, 8-byte stack; with setmaxnreg: C7506 ignored +// SASS: CALL.ABS.NOINC, then UTMALDG.2D; no USETMAXREG + +CTA form: tma_copy_2d_cta_test +// no address conversion: the shared symbols are the operands +cp.async.bulk.tensor.2d.shared::cta.global.tile + .mbarrier::complete_tx::bytes [TILE], [%rd2, {%r3, %r4}], [BAR]; +// ptxas sm_120a: 10 registers, no stack; with setmaxnreg: 168 registers +// no C7506; SASS: USETMAXREG.DEALLOC.CTAPOOL 0x28, UTMALDG.2D, +// USETMAXREG.TRY_ALLOC.CTAPOOL 0xe8; no CALL +What this proves: the cluster form converts the generic destination twice (to generic, then to the +cluster window) and ptxas on sm_120a adds the runtime CTA check whose call drops setmaxnreg. The CTA +form names the shared symbols directly and keeps register reallocation. On the 5090, tma_copy copies +4096 of 4096 values through the new kernel next to the unchanged ones, through both routes, with +memcheck and synccheck clean; ranks 1 to 5 are probed and assembled, and the cluster PTX is unchanged. + diff --git a/cuda-oxide/cuda-oxide-book/advanced/shared-memory-and-synchronization.md b/cuda-oxide/cuda-oxide-book/advanced/shared-memory-and-synchronization.md index 9a0c9f011..5871909e1 100644 --- a/cuda-oxide/cuda-oxide-book/advanced/shared-memory-and-synchronization.md +++ b/cuda-oxide/cuda-oxide-book/advanced/shared-memory-and-synchronization.md @@ -270,6 +270,69 @@ the strongest typed requirement over the kernel's `get`, `get_raw` and The declaration aligns the pool base only; keeping an `offset(byte_offset)` result aligned for `T` is still the caller's job. +### Typed shared pointers: `SharedPtr` + +`DynamicSharedArray::get()` and `SharedArray::as_mut_ptr()` return `*mut T`, +which in device code is a *generic* pointer: it may address global, local or +shared memory, including another block's shared memory in a cluster. Once +such a pointer is passed to a helper function, the compiler no longer knows it +points into this block's shared memory. + +`SharedPtr` is a raw pointer into the issuing block's shared memory +(`.shared::cta`) that carries that fact in its type. It lowers to an +`addrspace(3)` pointer wherever it goes, including helper parameters and +return values, struct fields and generic code, so instructions that take it +use their shared-memory forms without converting a generic address. + +```rust +use cuda_device::{DynamicSharedArray, SharedArray, SharedPtr, kernel, thread}; + +#[inline(never)] +unsafe fn fill(tile: SharedPtr, i: usize, value: f32) { + // Still a shared-memory pointer here, although `fill` is not inlined. + unsafe { tile.add(i).as_ptr().write(value) }; +} + +static mut TILE: SharedArray = SharedArray::UNINIT; + +#[kernel] +pub fn fill_tile() { + // SAFETY: TILE is a `static mut SharedArray` of this kernel. + let tile: SharedPtr = unsafe { SharedArray::shared_ptr(&raw mut TILE) }; + // The same type for dynamic shared memory, 1024 bytes into the pool. + let pool: SharedPtr = DynamicSharedArray::::shared_ptr_at(1024); + // SAFETY: the launch supplies more than 1024 dynamic shared bytes, and + // each thread writes its own slot of TILE. + unsafe { + fill(tile, thread::threadIdx_x() as usize, 1.0); + pool.as_ptr().write(0); + } +} +``` + +| Operation | Result | +|:---------------------------------------|:----------------------------------------------------| +| `unsafe SharedArray::shared_ptr(&raw mut S)` | `SharedPtr` to the first element of `S`, a `static mut` of this kernel | +| `DynamicSharedArray::::shared_ptr()` / `shared_ptr_at(bytes)` | Like `get()` / `offset(bytes)` | +| `unsafe SharedPtr::from_raw(p)` | Asserts that generic `p` points into this block's shared memory | +| `add`, `offset`, `byte_add`, `cast` | Pointer arithmetic that keeps the type | +| `as_ptr()` | The generic `*mut T`, for loads, stores and raw-pointer APIs | +| `addr_u32()` | The 32-bit `.shared::cta` address, for `_shared_u32` forms | + +`SharedPtr` cannot be a kernel parameter, bare or inside a struct or array: +the host has no shared-memory address to pass, and the compiler refuses the +signature. Pointers from `cluster::map_shared_rank` may name another block's +shared memory and are not `SharedPtr`s. + +Inside aggregates the usual shared-pointer limits apply. A union or a +`MaybeUninit>` fails closed under the union storage rules, and a +`repr(packed)` struct that holds one follows the packed-carrier limits; both +are listed in [Supported Features](../appendix/supported-features.md). On the +libNVVM route a shared pointer is 4 bytes wide, so a struct that mixes a +`SharedPtr` with a narrower field, such as `{ p: SharedPtr, x: u32 }`, is +not yet supported in arrays or in shared memory. That gap predates +`SharedPtr` and applies to `*mut Barrier` fields in the same way. + --- ## Synchronization: sync_threads() diff --git a/cuda-oxide/cuda-oxide-book/advanced/tensor-memory-accelerator.md b/cuda-oxide/cuda-oxide-book/advanced/tensor-memory-accelerator.md index add95f44b..f2397747b 100644 --- a/cuda-oxide/cuda-oxide-book/advanced/tensor-memory-accelerator.md +++ b/cuda-oxide/cuda-oxide-book/advanced/tensor-memory-accelerator.md @@ -119,6 +119,42 @@ The coordinates (`tile_x`, `tile_y`) specify which tile to copy, in units of the box dimensions encoded in the descriptor. The barrier pointer tells the TMA engine where to signal completion. +These functions address cluster shared memory: the destination may belong to +any CTA of the cluster, so the compiler converts it to the cluster window. +When the destination is the issuing CTA's own shared memory, +`cp_async_bulk_tensor_*_g2s_cta` emits the `.shared::cta` destination form +(PTX 8.6, sm_90+) without that conversion. Its destination is a +[`SharedPtr`](shared-memory-and-synchronization.md), +a CTA-shared pointer by type, so the locality holds wherever the pointer came +from, including through helper functions: + +```rust +// SAFETY: TILE and BAR are `static mut` shared-memory objects of this kernel, +// and the copy completes on BAR before any thread reads TILE. +unsafe { + cp_async_bulk_tensor_2d_g2s_cta( + SharedArray::shared_ptr(&raw mut TILE).cast::(), + tensor_map, tile_x, tile_y, + &raw mut BAR, + ); +} +``` + +On SM12x this form matters for performance: ptxas implements a non-multicast +`.shared::cluster` TMA load it cannot prove CTA-local with a runtime check and +a system call, which also makes it ignore `setmaxnreg`. + +```{figure} images/tma-cta-local.svg +:align: center +:width: 100% + +The two G2S destination forms. Left, the cluster form converts a generic `*mut u8` to the cluster +window, and on sm_120a ptxas adds a runtime CTA check whose call makes it ignore `setmaxnreg`. +Middle, `SharedPtr` keeps the shared address space in its type across helper parameters, struct +fields and return values. Right, `cp_async_bulk_tensor_2d_g2s_cta` takes a `SharedPtr` and emits +the `.shared::cta` form with no address conversion, so the kernel keeps `setmaxnreg` and its 168 registers. +``` + ### Only one thread issues the copy TMA copies are **not** collective operations across the block. Exactly diff --git a/cuda-oxide/cuda-oxide-book/appendix/api-quick-reference.md b/cuda-oxide/cuda-oxide-book/appendix/api-quick-reference.md index 4cd28a058..cc90fe89a 100644 --- a/cuda-oxide/cuda-oxide-book/appendix/api-quick-reference.md +++ b/cuda-oxide/cuda-oxide-book/appendix/api-quick-reference.md @@ -312,6 +312,7 @@ pub fn dynamic(data: &[f32]) { | `SharedArray` | Compile-time sized, block-scoped shared memory | | `SharedArray` | With 128-byte alignment (required for TMA destinations) | | `DynamicSharedArray` | Runtime-sized shared memory (set via `LaunchConfig`) | +| `SharedPtr` | Pointer into this block's shared memory, `addrspace(3)` by type | Both are `!Sync` — concurrent access requires explicit barriers. @@ -450,6 +451,7 @@ cp_async_bulk_commit_group(); | Function | Description | |:----------------------------------------------|:-------------------------------------| | `cp_async_bulk_tensor_{1..5}d_g2s(...)` | Global → shared async bulk copy | +| `cp_async_bulk_tensor_{1..5}d_g2s_cta(...)` | Global → this CTA's `SharedPtr` | | `cp_async_bulk_tensor_{1..5}d_s2g(...)` | Shared → global async bulk copy | | `cp_async_bulk_tensor_2d_g2s_multicast(...)` | Multicast to all CTAs in cluster | | `cp_async_bulk_commit_group()` | Commit outstanding copies | @@ -649,7 +651,7 @@ debug::prof_trigger::<7>(); // Nsight profiler trigger | `thread` | Thread/block IDs, `index_1d`, `sync_threads` | All | | `config` | Compile-time policies, shapes, tiles, atoms, layouts, memory spaces, and scopes | All | | `disjoint` | `DisjointSlice` — typed writes completed by a launch proof | All | -| `shared` | `SharedArray`, `DynamicSharedArray` | All | +| `shared` | `SharedArray`, `DynamicSharedArray`, `SharedPtr` | All | | `warp` | Shuffle, vote, match, lane/warp ID | All | | `atomic` | Scoped atomics (device/block/system) | sm_70+ | | `debug` | `clock64`, `trap`, `breakpoint`, `gpu_printf!` | All | diff --git a/cuda-oxide/cuda-oxide-book/appendix/supported-features.md b/cuda-oxide/cuda-oxide-book/appendix/supported-features.md index 9cc4d6774..f56825700 100644 --- a/cuda-oxide/cuda-oxide-book/appendix/supported-features.md +++ b/cuda-oxide/cuda-oxide-book/appendix/supported-features.md @@ -117,6 +117,13 @@ do not count toward that bound. Taking the in-place address of such a field (`u.ptr = p`, `&u.ptr`, `&mut u.ptr`) is refused with a diagnostic; build the union by value and read its fields by value. +`SharedPtr`, the typed shared pointer, follows the same limits: a union or a +`MaybeUninit>` fails closed under these union storage rules, a +`repr(packed)` struct that holds one follows the packed-carrier limits above, a +kernel parameter that holds one bare or inside a struct or array is refused, +and on the libNVVM route a struct that mixes one with a narrower field is not +yet supported in arrays or in shared memory. + Enum constants preserve payload relocations to device statics, including non-zero byte addends. This includes niche-encoded `Option<&T>` and direct-tagged enum layouts, both for direct thin-reference payloads and for diff --git a/cuda-oxide/cuda-oxide-book/compiler/mlir-dialects.md b/cuda-oxide/cuda-oxide-book/compiler/mlir-dialects.md index 96edf2387..7a22f93ff 100644 --- a/cuda-oxide/cuda-oxide-book/compiler/mlir-dialects.md +++ b/cuda-oxide/cuda-oxide-book/compiler/mlir-dialects.md @@ -303,11 +303,11 @@ they become `call` instructions to `@llvm.nvvm.*` intrinsics. ### Architecture Coverage -At catalog SHA-256 `31b52388` (the stamp in every `ops/generated/` file -header), the dialect holds 580 operations across 42 modules, and they come +At catalog SHA-256 `7e6b4b4b` (the stamp in every `ops/generated/` file +header), the dialect holds 585 operations across 42 modules, and they come from two different places. The split is the first thing to know about it, because it decides where -- and whether -- you would add one. If the header -stamp no longer starts with `31b52388`, the counts on this page predate the +stamp no longer starts with `7e6b4b4b`, the counts on this page predate the catalog you are reading. **Hand-written**, directly under `cuda-oxide/crates/dialect-nvvm/src/ops/`. These are the @@ -327,13 +327,13 @@ describe. There are seven modules and 26 operations: **Generated**, under `ops/generated/`, from `intrinsics/catalog.json` by `cuda-intrinsics-gen`. Every file there opens with `// @generated ... DO NOT EDIT.`, and editing one by hand is undone by the next generator run. This is -the large majority -- 35 modules and 554 operations, resolved from 1033 catalog +the large majority -- 35 modules and 559 operations, resolved from 1038 catalog entries, since several intrinsics can share one structural op: | Area | Modules | Ops | | :-------------------------- | :---------------------------------------------------------------------------- | --: | | Tensor Core Gen 5 + TMEM | `tcgen05` | 210 | -| Tensor Memory Accelerator | `tma` | 111 | +| Tensor Memory Accelerator | `tma` | 116 | | Special registers | `sreg` | 44 | | Packed (SIMD-in-register) | `packed_alu`, `packed_conversion`, `packed_atomic` | 55 | | Async copy and barriers | `cp_async`, `mbarrier_extended`, `mbarrier_basic`, `sync` | 35 | diff --git a/cuda-oxide/intrinsics/abi-v1.toml b/cuda-oxide/intrinsics/abi-v1.toml index c82c2cfa9..052304523 100644 --- a/cuda-oxide/intrinsics/abi-v1.toml +++ b/cuda-oxide/intrinsics/abi-v1.toml @@ -7201,3 +7201,38 @@ status = "active" catalog_id = "cvt_rn_satfinite_relu_e5m2x2_bf16x2" operation_key = "packed.convert.bf16x2.e5m2x2.nearest_even.satfinite.relu" raw_rust_signature = { safe = true, arguments = ["u32"], result = "u16" } + +[[entry]] +abi_id = "i1034" +status = "active" +catalog_id = "cp_async_bulk_tensor_1d_g2s_cta" +operation_key = "memory.copy.async.bulk.tensor.g2s.cta.tile.1d" +raw_rust_signature = { safe = false, arguments = ["*mut u8", "*const u8", "i32", "*mut u64"], result = "()" } + +[[entry]] +abi_id = "i1035" +status = "active" +catalog_id = "cp_async_bulk_tensor_2d_g2s_cta" +operation_key = "memory.copy.async.bulk.tensor.g2s.cta.tile.2d" +raw_rust_signature = { safe = false, arguments = ["*mut u8", "*const u8", "i32", "i32", "*mut u64"], result = "()" } + +[[entry]] +abi_id = "i1036" +status = "active" +catalog_id = "cp_async_bulk_tensor_3d_g2s_cta" +operation_key = "memory.copy.async.bulk.tensor.g2s.cta.tile.3d" +raw_rust_signature = { safe = false, arguments = ["*mut u8", "*const u8", "i32", "i32", "i32", "*mut u64"], result = "()" } + +[[entry]] +abi_id = "i1037" +status = "active" +catalog_id = "cp_async_bulk_tensor_4d_g2s_cta" +operation_key = "memory.copy.async.bulk.tensor.g2s.cta.tile.4d" +raw_rust_signature = { safe = false, arguments = ["*mut u8", "*const u8", "i32", "i32", "i32", "i32", "*mut u64"], result = "()" } + +[[entry]] +abi_id = "i1038" +status = "active" +catalog_id = "cp_async_bulk_tensor_5d_g2s_cta" +operation_key = "memory.copy.async.bulk.tensor.g2s.cta.tile.5d" +raw_rust_signature = { safe = false, arguments = ["*mut u8", "*const u8", "i32", "i32", "i32", "i32", "i32", "*mut u64"], result = "()" } diff --git a/cuda-oxide/intrinsics/catalog.json b/cuda-oxide/intrinsics/catalog.json index 52aea22d8..a4918e40b 100644 --- a/cuda-oxide/intrinsics/catalog.json +++ b/cuda-oxide/intrinsics/catalog.json @@ -11,8 +11,8 @@ }, "inputs": { "imported_sha256": "54d47b8e281dd953dc95fe16f834ff38164d9809171c9a1af097231b0bc45cab", - "overlay_sha256": "9a9ba81bc0e7a27a16ccf6a5254123c36ad98b47a8829bbc222cfb77173ff31e", - "abi_ledger_sha256": "0e8cec7fdd52ad28ff4365bea512101aef06f29ad7781dcde6b867ad93a0a5bc", + "overlay_sha256": "bdfaca84cf4ed51ee06ef7caf5ce1acc979849117f7fd7afcd65b8b8397bdb37", + "abi_ledger_sha256": "f4cf9e3a7cf9f40725d1e5a2645796167e9615c3c47d2998c67ab1c3952fb210", "evidence_sha256": [ "cuda-12.0-libnvvm-12.0.140-ex2-f16.json:7102ea7d567ff2d2eb75a0384148d2e5eb6e605c26a5b98c71371fa3c46adef6", "cuda-12.9-libnvvm-12.9.86-f32x2.json:bc5bb505677989affa9ac1044da772b4eb349700dc65e16f09fffffa1f5c493e", @@ -71,6 +71,7 @@ "cuda-13.3-libnvvm-13.3.73-packed-conversion-fp8-bf16x2.json:7cc62aa25f9d696eaefcaedf104f53da9d08f24385f18d1539308f34d56d38f8", "cuda-13.3-libnvvm-13.3.73-register-mma-mxf8f6f4-f32.json:e437bf84570f4df0a28a743a132e0b93f359f9431d1af03b443f324f62b715ba", "cuda-13.3-libnvvm-13.3.73-sparse-mma-fp8.json:8a57d207ffd510b8c1bb40643bdb828d3d22e5d1c5ab24e1bc1b697538e9bc21", + "cuda-13.3-libnvvm-13.3.73-tma-cta.json:352f3e82f0f1c190d99e12e717ef5357a05729521a1fbab100b40964bc9526e7", "cuda-13.3-libnvvm-13.3.73-tma-descriptor-control.json:ac44f485122404f9a900cfb94bff658160a6c23d9e2f8ef74bb751f044b848fc", "cuda-13.3-libnvvm-13.3.73-tma-reduce.json:f5df5bb51068966594660cb01b1b28c3b72cb004c456fcd8dff377ff0bee4fc0", "cuda-13.3-libnvvm-13.3.73-tma.json:8607f93605f215858a98075a0ea1394f2718d090538bcaf0a795e11b4cb66d47", @@ -129,6 +130,7 @@ "rust-llvm-23.1.3-aab11ee0-tcgen05-st.json:e559aa775e3e964f9889483e9d9d8c45b46628931ff78050c46f33e448f69e86", "rust-llvm-23.1.3-aab11ee0-tcgen05.json:b6cf11d81c0880877ac5ad8958231222c728d5ab0243ddc84f1db901fbc8608b", "rust-llvm-23.1.3-aab11ee0-threadfence.json:087343e0247e7f9571c74fe763c0d48fbd26e246a21a9f10e5466a9a4457f83e", + "rust-llvm-23.1.3-aab11ee0-tma-cta.json:5aa68833705b5a98b393c4c752739a93cb3360d8d1c2c130bc28e09e475a208c", "rust-llvm-23.1.3-aab11ee0-tma-descriptor-control.json:7b6deaedfb8f8ac4b6551c47701cdefd04867431ac4dc5c3fbaf8f8bc604c9e9", "rust-llvm-23.1.3-aab11ee0-tma-reduce.json:e472b96476cad1d406fdac6dd8209645357f068cbd545669e2f09a8c6a232b79", "rust-llvm-23.1.3-aab11ee0-tma.json:97fb372aa7c6966bd01667da32d9d0857e385776131cb548f7d70f73b52718bc", @@ -18551,6 +18553,274 @@ }, "summary": "Starts a TMA tile copy from global to cluster shared memory." }, + { + "id": "cp_async_bulk_tensor_1d_g2s_cta", + "operation_key": "memory.copy.async.bulk.tensor.g2s.cta.tile.1d", + "family": "tma", + "source": { + "kind": "llvm_imported", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_1d" + }, + "selections": [ + { + "source_record": "TMA_G2S_CTA_TILE_1D", + "asm": "cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0}}], [$mbar];", + "predicates": [ + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" + ], + "constraints": { + "immediate_bindings": [ + { + "argument_index": 5, + "value": 0 + } + ] + } + } + ], + "rust": { + "abi_id": "i1034", + "module": "tma", + "name": "cp_async_bulk_tensor_1d_g2s_cta", + "arguments": [ + "*mut u8", + "*const u8", + "i32", + "*mut u64" + ], + "result": "()", + "safe": false, + "must_use": false, + "safe_allowlist_reason": null, + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_1d_g2s_cta", + "compatibility_paths": [ + "cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta" + ] + }, + "dialect": { + "op_type": "CpAsyncBulkTensorG2sCtaTile1dOp", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_cta_tile_1d", + "operands": [ + "ptr", + "ptr", + "ptr", + "i32", + "i64" + ], + "results": [] + }, + "llvm": { + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.1d", + "arguments": [ + "shared_ptr", + "shared_ptr", + "ptr", + "i32", + "i64", + "i1" + ], + "results": [], + "properties": [ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly" + ], + "result_facts": { + "no_undef": false, + "range": null + } + }, + "semantics": { + "pure": false, + "memory": "read_write", + "convergent": true, + "execution_scope": "thread" + }, + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + }, + "ptx_result": "()", + "targets": "all", + "ptx_isa_version": "9.3", + "ptx_isa_section": "9.7.9.26.5 Asynchronous bulk tensor copy", + "ptx_isa_url": "https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor" + }, + "backend": { + "profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "status": "validated", + "target_triple": "nvptx64-nvidia-cuda", + "gpu_target": "sm_90", + "ptx_feature": "+ptx86" + }, + "backend_lowerings": [ + { + "backend": "llvm_nvptx", + "mechanism": "typed_nvvm", + "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + } + }, + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "artifact_path": null, + "build_id_prefix": null, + "status": "validated", + "stages": [ + { + "targets": [ + "sm_90" + ], + "representation": "separate typed LLVM declaration probe for the 1D CTA-local TMA G2S operation", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_1d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,007-byte source had SHA-256 253d3ea39f9d26f5000f83febb9f72a15af2bed6ad0f668ab206976ce4c6d6ba; the 2,640-byte bitcode had SHA-256 9f42944557579c5c9c4e38716cce52607a07c51e5f2272393948fdc40cfb8961; the 1,234-byte canonical IR had SHA-256 835c2efc3b53d5333ec6a8260fcad29d5d0ce889194d0523feb57f7b22281fff.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx86" + ], + "representation": "typed LLVM call for the 1D CTA-local TMA G2S operation", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,165-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 b0b4469f60fea3342f58e26eb4698834c6cb0714aac17fd3ffaed79cbd4f226a.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx86" + ], + "representation": "LLVM-produced PTX retaining the 1D CTA-local TMA G2S instruction", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", + "artifact_kind": "cubin", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "backend": "lib_nvvm", + "mechanism": "inline_ptx", + "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma-cta", + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + } + }, + "version": "CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0", + "sha256": "5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc", + "artifact_path": null, + "build_id_prefix": "88b802", + "status": "validated", + "stages": [ + { + "targets": [ + "compute_90", + "ptx93" + ], + "representation": "one NVVM IR 2.0 module containing the five CTA-local TMA G2S inline-PTX forms", + "stage": "backend_codegen", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx93" + ], + "representation": "linked cubin retaining five UTMALDG instructions", + "stage": "device_link", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions.", + "artifact_kind": "cubin", + "tool_path": "libnvJitLink.so.13.3.33", + "tool_version": "CUDA 13.3 nvJitLink 13.3.33", + "tool_sha256": "f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa" + } + ] + } + ], + "tma": { + "operation": "g2s_cta_tile1d", + "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", + "runtime_validation": "unexecuted" + }, + "lowering": "generated_tma", + "expected_ptx": { + "mnemonic": "cp", + "modifiers": [ + "async", + "bulk", + "tensor", + "1d", + "shared::cta", + "global", + "tile", + "mbarrier::complete_tx::bytes" + ], + "operands": [ + { + "kind": "address" + }, + { + "kind": "address" + }, + { + "kind": "address" + } + ] + }, + "summary": "Starts a TMA tile copy from global to the issuing CTA's shared memory." + }, { "id": "cp_async_bulk_tensor_1d_s2g", "operation_key": "memory.copy.async.bulk.tensor.s2g.tile.1d", @@ -19095,29 +19365,25 @@ "summary": "Starts a TMA tile copy from global to cluster shared memory." }, { - "id": "cp_async_bulk_tensor_2d_g2s_multicast", - "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.2d.multicast", + "id": "cp_async_bulk_tensor_2d_g2s_cta", + "operation_key": "memory.copy.async.bulk.tensor.g2s.cta.tile.2d", "family": "tma", "source": { "kind": "llvm_imported", - "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_2d" + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_2d" }, "selections": [ { - "source_record": "TMA_G2S_TILE_CG0_2D_MC", - "asm": "cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster$cg [$dst], [$tmap, {{$d0, $d1}}], [$mbar], $mc;", + "source_record": "TMA_G2S_CTA_TILE_2D", + "asm": "cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1}}], [$mbar];", "predicates": [ - "Subtarget->getPTXVersion() >= 80", - "Subtarget->getSmVersion() >= 90" + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" ], "constraints": { "immediate_bindings": [ { - "argument_index": 7, - "value": -1 - }, - { - "argument_index": 8, + "argument_index": 6, "value": 0 } ] @@ -19125,7 +19391,282 @@ } ], "rust": { - "abi_id": "i0330", + "abi_id": "i1035", + "module": "tma", + "name": "cp_async_bulk_tensor_2d_g2s_cta", + "arguments": [ + "*mut u8", + "*const u8", + "i32", + "i32", + "*mut u64" + ], + "result": "()", + "safe": false, + "must_use": false, + "safe_allowlist_reason": null, + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_cta", + "compatibility_paths": [ + "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta" + ] + }, + "dialect": { + "op_type": "CpAsyncBulkTensorG2sCtaTile2dOp", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_cta_tile_2d", + "operands": [ + "ptr", + "ptr", + "ptr", + "i32", + "i32", + "i64" + ], + "results": [] + }, + "llvm": { + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d", + "arguments": [ + "shared_ptr", + "shared_ptr", + "ptr", + "i32", + "i32", + "i64", + "i1" + ], + "results": [], + "properties": [ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "WriteOnly" + ], + "result_facts": { + "no_undef": false, + "range": null + } + }, + "semantics": { + "pure": false, + "memory": "read_write", + "convergent": true, + "execution_scope": "thread" + }, + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + }, + "ptx_result": "()", + "targets": "all", + "ptx_isa_version": "9.3", + "ptx_isa_section": "9.7.9.26.5 Asynchronous bulk tensor copy", + "ptx_isa_url": "https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor" + }, + "backend": { + "profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "status": "validated", + "target_triple": "nvptx64-nvidia-cuda", + "gpu_target": "sm_90", + "ptx_feature": "+ptx86" + }, + "backend_lowerings": [ + { + "backend": "llvm_nvptx", + "mechanism": "typed_nvvm", + "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + } + }, + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "artifact_path": null, + "build_id_prefix": null, + "status": "validated", + "stages": [ + { + "targets": [ + "sm_90" + ], + "representation": "separate typed LLVM declaration probe for the 2D CTA-local TMA G2S operation", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_2d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,038-byte source had SHA-256 f0d47c1b6670854dfddce5bc45206d8f4ad16e9a05236dd4e1d4036b518666ea; the 2,648-byte bitcode had SHA-256 6ab53bfd2185eb8b6d7f7ffb3793d72cb7b48a07f61846a5317604700bdd3225; the 1,265-byte canonical IR had SHA-256 2c00613205683c093c5002f0f7347d3b2d6b82e46c5bc00072fb5a781f7d9d44.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx86" + ], + "representation": "typed LLVM call for the 2D CTA-local TMA G2S operation", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,299-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 59fa9746f48c669cbff59faffea6039a796055f2721b8647a6331d78b5a26521.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx86" + ], + "representation": "LLVM-produced PTX retaining the 2D CTA-local TMA G2S instruction", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", + "artifact_kind": "cubin", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "backend": "lib_nvvm", + "mechanism": "inline_ptx", + "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma-cta", + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + } + }, + "version": "CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0", + "sha256": "5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc", + "artifact_path": null, + "build_id_prefix": "88b802", + "status": "validated", + "stages": [ + { + "targets": [ + "compute_90", + "ptx93" + ], + "representation": "one NVVM IR 2.0 module containing the five CTA-local TMA G2S inline-PTX forms", + "stage": "backend_codegen", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx93" + ], + "representation": "linked cubin retaining five UTMALDG instructions", + "stage": "device_link", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions.", + "artifact_kind": "cubin", + "tool_path": "libnvJitLink.so.13.3.33", + "tool_version": "CUDA 13.3 nvJitLink 13.3.33", + "tool_sha256": "f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa" + } + ] + } + ], + "tma": { + "operation": "g2s_cta_tile2d", + "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", + "runtime_validation": "unexecuted" + }, + "lowering": "generated_tma", + "expected_ptx": { + "mnemonic": "cp", + "modifiers": [ + "async", + "bulk", + "tensor", + "2d", + "shared::cta", + "global", + "tile", + "mbarrier::complete_tx::bytes" + ], + "operands": [ + { + "kind": "address" + }, + { + "kind": "address" + }, + { + "kind": "address" + } + ] + }, + "summary": "Starts a TMA tile copy from global to the issuing CTA's shared memory." + }, + { + "id": "cp_async_bulk_tensor_2d_g2s_multicast", + "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.2d.multicast", + "family": "tma", + "source": { + "kind": "llvm_imported", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_2d" + }, + "selections": [ + { + "source_record": "TMA_G2S_TILE_CG0_2D_MC", + "asm": "cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster$cg [$dst], [$tmap, {{$d0, $d1}}], [$mbar], $mc;", + "predicates": [ + "Subtarget->getPTXVersion() >= 80", + "Subtarget->getSmVersion() >= 90" + ], + "constraints": { + "immediate_bindings": [ + { + "argument_index": 7, + "value": -1 + }, + { + "argument_index": 8, + "value": 0 + } + ] + } + } + ], + "rust": { + "abi_id": "i0330", "module": "tma", "name": "cp_async_bulk_tensor_2d_g2s_multicast", "arguments": [ @@ -19161,13 +19702,889 @@ "results": [] }, "llvm": { - "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d", + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d", + "arguments": [ + "shared_cluster_ptr", + "shared_ptr", + "ptr", + "i32", + "i32", + "i16", + "i64", + "i1", + "i1", + "i32" + ], + "results": [], + "properties": [ + "ImmArg", + "ImmArg", + "ImmArg", + "IntrConvergent", + "Range", + "ReadOnly", + "WriteOnly" + ], + "result_facts": { + "no_undef": false, + "range": null + } + }, + "semantics": { + "pure": false, + "memory": "read_write", + "convergent": true, + "execution_scope": "thread" + }, + "target": { + "minimum_ptx": "8.0", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + }, + "ptx_result": "()", + "targets": "all", + "ptx_isa_version": "9.3", + "ptx_isa_section": "9.7.9.26.5 Asynchronous bulk tensor copy", + "ptx_isa_url": "https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor" + }, + "backend": { + "profile": "rust-llvm-23.1.3-aab11ee0-tma", + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "status": "validated", + "target_triple": "nvptx64-nvidia-cuda", + "gpu_target": "sm_90", + "ptx_feature": "+ptx80" + }, + "backend_lowerings": [ + { + "backend": "llvm_nvptx", + "mechanism": "typed_nvvm", + "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma", + "target": { + "minimum_ptx": "8.0", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + } + }, + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "artifact_path": null, + "build_id_prefix": null, + "status": "validated", + "stages": [ + { + "targets": [ + "sm_90" + ], + "representation": "separate typed LLVM declaration probes", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx80" + ], + "representation": "separate probes for the recorded lowering mechanism", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): the exact reviewed PTX instruction shapes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 418eaf16785cbdd0e5b8fb5d886e2768f4347b8dda669cd924ad0cd857e7e08a.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx80" + ], + "representation": "separate PTX outputs assembled for the recorded target", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): terminal PTX assembly passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 45075184f52be336a8f24e8232251aec1a5890d9a6f0d35cbef2602dca1dafaa.", + "artifact_kind": "cubin", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "backend": "lib_nvvm", + "mechanism": "inline_ptx", + "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma", + "target": { + "minimum_ptx": "8.0", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + } + }, + "version": "CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0", + "sha256": "5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc", + "artifact_path": null, + "build_id_prefix": "88b802", + "status": "validated", + "stages": [ + { + "targets": [ + "compute_90", + "ptx93" + ], + "representation": "14 per-operation NVVM IR 2.0 modules covering the non-CTA-pair inline-PTX forms", + "stage": "backend_codegen", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx93" + ], + "representation": "14 linked cubins covering the non-CTA-pair TMA instructions", + "stage": "device_link", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303.", + "artifact_kind": "cubin", + "tool_path": "libnvJitLink.so.13.3.33", + "tool_version": "CUDA 13.3 nvJitLink 13.3.33", + "tool_sha256": "f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa" + } + ] + } + ], + "tma": { + "operation": "g2s_tile2d_multicast", + "adapter": "g2s_pointers_coordinates_barrier_mask_inject_defaults", + "runtime_validation": "unexecuted" + }, + "lowering": "generated_tma", + "expected_ptx": { + "mnemonic": "cp", + "modifiers": [ + "async", + "bulk", + "tensor", + "2d", + "shared::cluster", + "global", + "tile", + "mbarrier::complete_tx::bytes", + "multicast::cluster" + ], + "operands": [ + { + "kind": "address" + }, + { + "kind": "address" + }, + { + "kind": "address" + }, + { + "kind": "register" + } + ] + }, + "summary": "Starts a multicast TMA tile copy from global to cluster shared memory." + }, + { + "id": "cp_async_bulk_tensor_2d_g2s_multicast_cg2", + "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.2d.multicast.cta_group_2", + "family": "tma", + "source": { + "kind": "llvm_imported", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_2d" + }, + "selections": [ + { + "source_record": "TMA_G2S_TILE_2D_MC", + "asm": "cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster$cg [$dst], [$tmap, {{$d0, $d1}}], [$mbar], $mc;", + "predicates": [ + "Subtarget->hasTMABlackwellSupport()" + ], + "constraints": { + "immediate_bindings": [ + { + "argument_index": 7, + "value": -1 + }, + { + "argument_index": 8, + "value": 0 + } + ] + } + } + ], + "rust": { + "abi_id": "i0331", + "module": "tma", + "name": "cp_async_bulk_tensor_2d_g2s_multicast_cg2", + "arguments": [ + "*mut u8", + "*const u8", + "i32", + "i32", + "*mut u64", + "u16" + ], + "result": "()", + "safe": false, + "must_use": false, + "safe_allowlist_reason": null, + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0331", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_multicast_cg2", + "compatibility_paths": [ + "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_multicast_cg2" + ] + }, + "dialect": { + "op_type": "CpAsyncBulkTensorG2sTile2dMulticastCg2Op", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_tile_2d_multicast_cg2", + "operands": [ + "ptr", + "ptr", + "ptr", + "i32", + "i32", + "i16", + "i64" + ], + "results": [] + }, + "llvm": { + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d", + "arguments": [ + "shared_cluster_ptr", + "shared_ptr", + "ptr", + "i32", + "i32", + "i16", + "i64", + "i1", + "i1", + "i32" + ], + "results": [], + "properties": [ + "ImmArg", + "ImmArg", + "ImmArg", + "IntrConvergent", + "Range", + "ReadOnly", + "WriteOnly" + ], + "result_facts": { + "no_undef": false, + "range": null + } + }, + "semantics": { + "pure": false, + "memory": "read_write", + "convergent": true, + "execution_scope": "thread" + }, + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "exact_architecture", + "sm": 100 + }, + { + "kind": "exact_architecture", + "sm": 101 + }, + { + "kind": "exact_architecture", + "sm": 103 + }, + { + "kind": "exact_architecture", + "sm": 110 + } + ] + }, + "ptx_result": "()", + "targets": "sm_100a|sm_101a|sm_103a|sm_110a", + "ptx_isa_version": "9.3", + "ptx_isa_section": "9.7.9.26.5 Asynchronous bulk tensor copy", + "ptx_isa_url": "https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor" + }, + "backend": { + "profile": "rust-llvm-23.1.3-aab11ee0-tma", + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "status": "validated", + "target_triple": "nvptx64-nvidia-cuda", + "gpu_target": "sm_100a", + "ptx_feature": "+ptx86" + }, + "backend_lowerings": [ + { + "backend": "llvm_nvptx", + "mechanism": "typed_nvvm", + "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma", + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "exact_architecture", + "sm": 100 + }, + { + "kind": "exact_architecture", + "sm": 101 + }, + { + "kind": "exact_architecture", + "sm": 103 + }, + { + "kind": "exact_architecture", + "sm": 110 + } + ] + } + }, + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "artifact_path": null, + "build_id_prefix": null, + "status": "validated", + "stages": [ + { + "targets": [ + "sm_100a" + ], + "representation": "separate typed LLVM declaration probes", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 1 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 065ffbad7cc3f4a868f3de9e7f86d3a4fb1b0ea7047e1ddcdd578192e1fe101c.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_100a", + "ptx86" + ], + "representation": "separate probes for the recorded lowering mechanism", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 1 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): the exact reviewed PTX instruction shapes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 c6f52db1ccb491246b76348fb43a73fc961823d087d80432a2cec08dc36439f1.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_100a", + "ptx86" + ], + "representation": "separate PTX outputs assembled for the recorded target", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 1 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): terminal PTX assembly passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 eea1fea3e3abbce7a6d377fafd8ad8e2c8fec998a990afbd664b7ce4f3856dff.", + "artifact_kind": "cubin", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "backend": "lib_nvvm", + "mechanism": "inline_ptx", + "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma", + "target": { + "minimum_ptx": "8.6", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "exact_architecture", + "sm": 100 + }, + { + "kind": "exact_architecture", + "sm": 101 + }, + { + "kind": "exact_architecture", + "sm": 103 + }, + { + "kind": "exact_architecture", + "sm": 110 + } + ] + } + }, + "version": "CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0", + "sha256": "5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc", + "artifact_path": null, + "build_id_prefix": "88b802", + "status": "validated", + "stages": [ + { + "targets": [ + "compute_100a", + "ptx93" + ], + "representation": "one NVVM IR 2.0 module containing the exact CTA-pair inline-PTX form", + "stage": "backend_codegen", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3.73 libNVVM retained the exact cta_group::2 instruction with no extern call. The source IR SHA-256 is be32e3ad9cf01c97078b9a6799a994afc5e1c1264fd0ad0831b2e98f6f2275ca, LTOIR SHA-256 is ebc556f17c97cd4da6eb87512373c3d93bd44ba3faf68db41206f934a22a60ac, and linked PTX SHA-256 is 8b48f700abf6f55e3dc13443bf1ffe7a45f639cd16f683443cfbf5c84feacc38.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_100a", + "ptx93" + ], + "representation": "linked cubin retaining the CTA-pair TMA instruction", + "stage": "device_link", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3 nvJitLink produced a valid sm_100a CUDA ELF cubin with SHA-256 07a9b441996fc5ba0b8a05aebf0ce67218f0ea8e75f45e4c652a0778207a5343.", + "artifact_kind": "cubin", + "tool_path": "libnvJitLink.so.13.3.33", + "tool_version": "CUDA 13.3 nvJitLink 13.3.33", + "tool_sha256": "f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa" + } + ] + } + ], + "tma": { + "operation": "g2s_tile2d_multicast_cg2", + "adapter": "g2s_pointers_coordinates_barrier_mask_inject_defaults", + "runtime_validation": "unexecuted" + }, + "lowering": "generated_tma", + "expected_ptx": { + "mnemonic": "cp", + "modifiers": [ + "async", + "bulk", + "tensor", + "2d", + "shared::cluster", + "global", + "tile", + "mbarrier::complete_tx::bytes", + "multicast::cluster", + "cta_group::2" + ], + "operands": [ + { + "kind": "address" + }, + { + "kind": "address" + }, + { + "kind": "address" + }, + { + "kind": "register" + } + ] + }, + "summary": "Starts a multicast TMA tile copy from global to cluster shared memory." + }, + { + "id": "cp_async_bulk_tensor_2d_s2g", + "operation_key": "memory.copy.async.bulk.tensor.s2g.tile.2d", + "family": "tma", + "source": { + "kind": "llvm_imported", + "source_record": "int_nvvm_cp_async_bulk_tensor_s2g_tile_2d" + }, + "selections": [ + { + "source_record": "TMA_TENSOR_S2G_TILE_2D", + "asm": "cp.async.bulk.tensor.2d.global.shared::cta.tile.bulk_group [$tmap, {{$d0, $d1}}], [$src];", + "predicates": [ + "Subtarget->getPTXVersion() >= 80", + "Subtarget->getSmVersion() >= 90" + ], + "constraints": { + "immediate_bindings": [ + { + "argument_index": 5, + "value": 0 + } + ] + } + } + ], + "rust": { + "abi_id": "i0336", + "module": "tma", + "name": "cp_async_bulk_tensor_2d_s2g", + "arguments": [ + "*const u8", + "*const u8", + "i32", + "i32" + ], + "result": "()", + "safe": false, + "must_use": false, + "safe_allowlist_reason": null, + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0336", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_s2g", + "compatibility_paths": [ + "cuda_device::tma::cp_async_bulk_tensor_2d_s2g" + ] + }, + "dialect": { + "op_type": "CpAsyncBulkTensorS2gTile2dOp", + "op_name": "nvvm.cp_async_bulk_tensor_s2g_tile_2d", + "operands": [ + "ptr", + "ptr", + "i32", + "i32" + ], + "results": [] + }, + "llvm": { + "symbol": "llvm.nvvm.cp.async.bulk.tensor.s2g.tile.2d", + "arguments": [ + "shared_ptr", + "ptr", + "i32", + "i32", + "i64", + "i1" + ], + "results": [], + "properties": [ + "ImmArg", + "IntrConvergent", + "ReadOnly", + "ReadOnly" + ], + "result_facts": { + "no_undef": false, + "range": null + } + }, + "semantics": { + "pure": false, + "memory": "read_write", + "convergent": true, + "execution_scope": "thread" + }, + "target": { + "minimum_ptx": "8.0", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + }, + "ptx_result": "()", + "targets": "all", + "ptx_isa_version": "9.3", + "ptx_isa_section": "9.7.9.26.5 Asynchronous bulk tensor copy", + "ptx_isa_url": "https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor" + }, + "backend": { + "profile": "rust-llvm-23.1.3-aab11ee0-tma", + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "status": "validated", + "target_triple": "nvptx64-nvidia-cuda", + "gpu_target": "sm_90", + "ptx_feature": "+ptx80" + }, + "backend_lowerings": [ + { + "backend": "llvm_nvptx", + "mechanism": "typed_nvvm", + "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma", + "target": { + "minimum_ptx": "8.0", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + } + }, + "version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "artifact_path": null, + "build_id_prefix": null, + "status": "validated", + "stages": [ + { + "targets": [ + "sm_90" + ], + "representation": "separate typed LLVM declaration probes", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx80" + ], + "representation": "separate probes for the recorded lowering mechanism", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): the exact reviewed PTX instruction shapes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 418eaf16785cbdd0e5b8fb5d886e2768f4347b8dda669cd924ad0cd857e7e08a.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx80" + ], + "representation": "separate PTX outputs assembled for the recorded target", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): terminal PTX assembly passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 45075184f52be336a8f24e8232251aec1a5890d9a6f0d35cbef2602dca1dafaa.", + "artifact_kind": "cubin", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "backend": "lib_nvvm", + "mechanism": "inline_ptx", + "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma", + "target": { + "minimum_ptx": "8.0", + "hardware": { + "kind": "any_of", + "alternatives": [ + { + "kind": "minimum_sm", + "sm": 90 + } + ] + } + }, + "version": "CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0", + "sha256": "5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc", + "artifact_path": null, + "build_id_prefix": "88b802", + "status": "validated", + "stages": [ + { + "targets": [ + "compute_90", + "ptx93" + ], + "representation": "14 per-operation NVVM IR 2.0 modules covering the non-CTA-pair inline-PTX forms", + "stage": "backend_codegen", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX.", + "artifact_kind": null, + "tool_path": null, + "tool_version": null, + "tool_sha256": null + }, + { + "targets": [ + "sm_90", + "ptx93" + ], + "representation": "14 linked cubins covering the non-CTA-pair TMA instructions", + "stage": "device_link", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303.", + "artifact_kind": "cubin", + "tool_path": "libnvJitLink.so.13.3.33", + "tool_version": "CUDA 13.3 nvJitLink 13.3.33", + "tool_sha256": "f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa" + } + ] + } + ], + "tma": { + "operation": "s2g_tile2d", + "adapter": "s2g_pointers_coordinates_inject_defaults", + "runtime_validation": "unexecuted" + }, + "lowering": "generated_tma", + "expected_ptx": { + "mnemonic": "cp", + "modifiers": [ + "async", + "bulk", + "tensor", + "2d", + "global", + "shared::cta", + "tile", + "bulk_group" + ], + "operands": [ + { + "kind": "address" + }, + { + "kind": "address" + } + ] + }, + "summary": "Starts a TMA tile copy from shared to global memory." + }, + { + "id": "cp_async_bulk_tensor_3d_g2s", + "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.3d", + "family": "tma", + "source": { + "kind": "llvm_imported", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_3d" + }, + "selections": [ + { + "source_record": "TMA_G2S_TILE_CG0_3D", + "asm": "cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2}}], [$mbar];", + "predicates": [ + "Subtarget->getPTXVersion() >= 80", + "Subtarget->getSmVersion() >= 90" + ], + "constraints": { + "immediate_bindings": [ + { + "argument_index": 8, + "value": 0 + }, + { + "argument_index": 9, + "value": 0 + } + ] + } + } + ], + "rust": { + "abi_id": "i0332", + "module": "tma", + "name": "cp_async_bulk_tensor_3d_g2s", + "arguments": [ + "*mut u8", + "*const u8", + "i32", + "i32", + "i32", + "*mut u64" + ], + "result": "()", + "safe": false, + "must_use": false, + "safe_allowlist_reason": null, + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0332", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s", + "compatibility_paths": [ + "cuda_device::tma::cp_async_bulk_tensor_3d_g2s" + ] + }, + "dialect": { + "op_type": "CpAsyncBulkTensorG2sTile3dOp", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_tile_3d", + "operands": [ + "ptr", + "ptr", + "ptr", + "i32", + "i32", + "i32", + "i16", + "i64" + ], + "results": [] + }, + "llvm": { + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.3d", "arguments": [ "shared_cluster_ptr", "shared_ptr", "ptr", "i32", "i32", + "i32", "i16", "i64", "i1", @@ -19176,11 +20593,11 @@ ], "results": [], "properties": [ - "ImmArg", + "ImmArg", "ImmArg", "ImmArg", "IntrConvergent", - "Range", + "Range", "ReadOnly", "WriteOnly" ], @@ -19346,8 +20763,8 @@ } ], "tma": { - "operation": "g2s_tile2d_multicast", - "adapter": "g2s_pointers_coordinates_barrier_mask_inject_defaults", + "operation": "g2s_tile3d", + "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", "runtime_validation": "unexecuted" }, "lowering": "generated_tma", @@ -19357,12 +20774,11 @@ "async", "bulk", "tensor", - "2d", + "3d", "shared::cluster", "global", "tile", - "mbarrier::complete_tx::bytes", - "multicast::cluster" + "mbarrier::complete_tx::bytes" ], "operands": [ { @@ -19373,37 +20789,31 @@ }, { "kind": "address" - }, - { - "kind": "register" } ] }, - "summary": "Starts a multicast TMA tile copy from global to cluster shared memory." + "summary": "Starts a TMA tile copy from global to cluster shared memory." }, { - "id": "cp_async_bulk_tensor_2d_g2s_multicast_cg2", - "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.2d.multicast.cta_group_2", + "id": "cp_async_bulk_tensor_3d_g2s_cta", + "operation_key": "memory.copy.async.bulk.tensor.g2s.cta.tile.3d", "family": "tma", "source": { "kind": "llvm_imported", - "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_2d" + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_3d" }, "selections": [ { - "source_record": "TMA_G2S_TILE_2D_MC", - "asm": "cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster$cg [$dst], [$tmap, {{$d0, $d1}}], [$mbar], $mc;", + "source_record": "TMA_G2S_CTA_TILE_3D", + "asm": "cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2}}], [$mbar];", "predicates": [ - "Subtarget->hasTMABlackwellSupport()" + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" ], "constraints": { "immediate_bindings": [ { "argument_index": 7, - "value": -1 - }, - { - "argument_index": 8, "value": 0 } ] @@ -19411,62 +20821,57 @@ } ], "rust": { - "abi_id": "i0331", + "abi_id": "i1036", "module": "tma", - "name": "cp_async_bulk_tensor_2d_g2s_multicast_cg2", + "name": "cp_async_bulk_tensor_3d_g2s_cta", "arguments": [ "*mut u8", "*const u8", "i32", "i32", - "*mut u64", - "u16" + "i32", + "*mut u64" ], "result": "()", "safe": false, "must_use": false, "safe_allowlist_reason": null, - "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0331", - "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_multicast_cg2", + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s_cta", "compatibility_paths": [ - "cuda_device::tma::cp_async_bulk_tensor_2d_g2s_multicast_cg2" + "cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta" ] }, "dialect": { - "op_type": "CpAsyncBulkTensorG2sTile2dMulticastCg2Op", - "op_name": "nvvm.cp_async_bulk_tensor_g2s_tile_2d_multicast_cg2", + "op_type": "CpAsyncBulkTensorG2sCtaTile3dOp", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_cta_tile_3d", "operands": [ "ptr", "ptr", "ptr", "i32", "i32", - "i16", + "i32", "i64" ], "results": [] }, "llvm": { - "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d", + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.3d", "arguments": [ - "shared_cluster_ptr", + "shared_ptr", "shared_ptr", "ptr", "i32", "i32", - "i16", + "i32", "i64", - "i1", - "i1", - "i32" + "i1" ], "results": [], "properties": [ "ImmArg", - "ImmArg", - "ImmArg", "IntrConvergent", - "Range", "ReadOnly", "WriteOnly" ], @@ -19487,63 +20892,39 @@ "kind": "any_of", "alternatives": [ { - "kind": "exact_architecture", - "sm": 100 - }, - { - "kind": "exact_architecture", - "sm": 101 - }, - { - "kind": "exact_architecture", - "sm": 103 - }, - { - "kind": "exact_architecture", - "sm": 110 + "kind": "minimum_sm", + "sm": 90 } ] }, "ptx_result": "()", - "targets": "sm_100a|sm_101a|sm_103a|sm_110a", + "targets": "all", "ptx_isa_version": "9.3", "ptx_isa_section": "9.7.9.26.5 Asynchronous bulk tensor copy", "ptx_isa_url": "https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor" }, "backend": { - "profile": "rust-llvm-23.1.3-aab11ee0-tma", + "profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", "version": "LLVM version 23.1.3-rust-1.101.0-nightly", "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", "status": "validated", "target_triple": "nvptx64-nvidia-cuda", - "gpu_target": "sm_100a", + "gpu_target": "sm_90", "ptx_feature": "+ptx86" }, "backend_lowerings": [ { "backend": "llvm_nvptx", "mechanism": "typed_nvvm", - "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma", + "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", "target": { "minimum_ptx": "8.6", "hardware": { "kind": "any_of", "alternatives": [ { - "kind": "exact_architecture", - "sm": 100 - }, - { - "kind": "exact_architecture", - "sm": 101 - }, - { - "kind": "exact_architecture", - "sm": 103 - }, - { - "kind": "exact_architecture", - "sm": 110 + "kind": "minimum_sm", + "sm": 90 } ] } @@ -19556,13 +20937,13 @@ "stages": [ { "targets": [ - "sm_100a" + "sm_90" ], - "representation": "separate typed LLVM declaration probes", + "representation": "separate typed LLVM declaration probe for the 3D CTA-local TMA G2S operation", "stage": "declaration_canonicalization", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 1 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 065ffbad7cc3f4a868f3de9e7f86d3a4fb1b0ea7047e1ddcdd578192e1fe101c.", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_3d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,069-byte source had SHA-256 159ab50347ad4f81c8a110df7de35db35da9eb099938874523e5ec10ea867c8c; the 2,656-byte bitcode had SHA-256 3b5f19c93cffadf279efad33b501519f09cacbc660784f195ab3d007b5bca854; the 1,296-byte canonical IR had SHA-256 223dd701edae4ecf145ca18afc659dd5a50785cf8e492f6ba5c7b7f8299dc030.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -19570,14 +20951,14 @@ }, { "targets": [ - "sm_100a", + "sm_90", "ptx86" ], - "representation": "separate probes for the recorded lowering mechanism", + "representation": "typed LLVM call for the 3D CTA-local TMA G2S operation", "stage": "backend_codegen", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 1 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): the exact reviewed PTX instruction shapes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 c6f52db1ccb491246b76348fb43a73fc961823d087d80432a2cec08dc36439f1.", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,433-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 e9bab5f8e2c2a697ca435f86eaff468984a03470ae4bf28f975c2c89fc6676b9.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -19585,14 +20966,14 @@ }, { "targets": [ - "sm_100a", + "sm_90", "ptx86" ], - "representation": "separate PTX outputs assembled for the recorded target", + "representation": "LLVM-produced PTX retaining the 3D CTA-local TMA G2S instruction", "stage": "ptx_assembly", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 1 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): terminal PTX assembly passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 eea1fea3e3abbce7a6d377fafd8ad8e2c8fec998a990afbd664b7ce4f3856dff.", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", "artifact_kind": "cubin", "tool_path": "/usr/local/cuda-13.3/bin/ptxas", "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", @@ -19603,27 +20984,15 @@ { "backend": "lib_nvvm", "mechanism": "inline_ptx", - "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma", + "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma-cta", "target": { "minimum_ptx": "8.6", "hardware": { "kind": "any_of", "alternatives": [ { - "kind": "exact_architecture", - "sm": 100 - }, - { - "kind": "exact_architecture", - "sm": 101 - }, - { - "kind": "exact_architecture", - "sm": 103 - }, - { - "kind": "exact_architecture", - "sm": 110 + "kind": "minimum_sm", + "sm": 90 } ] } @@ -19636,14 +21005,14 @@ "stages": [ { "targets": [ - "compute_100a", + "compute_90", "ptx93" ], - "representation": "one NVVM IR 2.0 module containing the exact CTA-pair inline-PTX form", + "representation": "one NVVM IR 2.0 module containing the five CTA-local TMA G2S inline-PTX forms", "stage": "backend_codegen", "mechanism": "inline_ptx", "outcome": "succeeded", - "detail": "CUDA 13.3.73 libNVVM retained the exact cta_group::2 instruction with no extern call. The source IR SHA-256 is be32e3ad9cf01c97078b9a6799a994afc5e1c1264fd0ad0831b2e98f6f2275ca, LTOIR SHA-256 is ebc556f17c97cd4da6eb87512373c3d93bd44ba3faf68db41206f934a22a60ac, and linked PTX SHA-256 is 8b48f700abf6f55e3dc13443bf1ffe7a45f639cd16f683443cfbf5c84feacc38.", + "detail": "CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -19651,14 +21020,14 @@ }, { "targets": [ - "sm_100a", + "sm_90", "ptx93" ], - "representation": "linked cubin retaining the CTA-pair TMA instruction", + "representation": "linked cubin retaining five UTMALDG instructions", "stage": "device_link", "mechanism": "inline_ptx", "outcome": "succeeded", - "detail": "CUDA 13.3 nvJitLink produced a valid sm_100a CUDA ELF cubin with SHA-256 07a9b441996fc5ba0b8a05aebf0ce67218f0ea8e75f45e4c652a0778207a5343.", + "detail": "CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions.", "artifact_kind": "cubin", "tool_path": "libnvJitLink.so.13.3.33", "tool_version": "CUDA 13.3 nvJitLink 13.3.33", @@ -19668,8 +21037,8 @@ } ], "tma": { - "operation": "g2s_tile2d_multicast_cg2", - "adapter": "g2s_pointers_coordinates_barrier_mask_inject_defaults", + "operation": "g2s_cta_tile3d", + "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", "runtime_validation": "unexecuted" }, "lowering": "generated_tma", @@ -19679,13 +21048,11 @@ "async", "bulk", "tensor", - "2d", - "shared::cluster", + "3d", + "shared::cta", "global", "tile", - "mbarrier::complete_tx::bytes", - "multicast::cluster", - "cta_group::2" + "mbarrier::complete_tx::bytes" ], "operands": [ { @@ -19696,26 +21063,23 @@ }, { "kind": "address" - }, - { - "kind": "register" } ] }, - "summary": "Starts a multicast TMA tile copy from global to cluster shared memory." + "summary": "Starts a TMA tile copy from global to the issuing CTA's shared memory." }, { - "id": "cp_async_bulk_tensor_2d_s2g", - "operation_key": "memory.copy.async.bulk.tensor.s2g.tile.2d", + "id": "cp_async_bulk_tensor_3d_s2g", + "operation_key": "memory.copy.async.bulk.tensor.s2g.tile.3d", "family": "tma", "source": { "kind": "llvm_imported", - "source_record": "int_nvvm_cp_async_bulk_tensor_s2g_tile_2d" + "source_record": "int_nvvm_cp_async_bulk_tensor_s2g_tile_3d" }, "selections": [ { - "source_record": "TMA_TENSOR_S2G_TILE_2D", - "asm": "cp.async.bulk.tensor.2d.global.shared::cta.tile.bulk_group [$tmap, {{$d0, $d1}}], [$src];", + "source_record": "TMA_TENSOR_S2G_TILE_3D", + "asm": "cp.async.bulk.tensor.3d.global.shared::cta.tile.bulk_group [$tmap, {{$d0, $d1, $d2}}], [$src];", "predicates": [ "Subtarget->getPTXVersion() >= 80", "Subtarget->getSmVersion() >= 90" @@ -19723,7 +21087,7 @@ "constraints": { "immediate_bindings": [ { - "argument_index": 5, + "argument_index": 6, "value": 0 } ] @@ -19731,49 +21095,52 @@ } ], "rust": { - "abi_id": "i0336", + "abi_id": "i0337", "module": "tma", - "name": "cp_async_bulk_tensor_2d_s2g", + "name": "cp_async_bulk_tensor_3d_s2g", "arguments": [ "*const u8", "*const u8", "i32", + "i32", "i32" ], "result": "()", "safe": false, "must_use": false, "safe_allowlist_reason": null, - "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0336", - "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_2d_s2g", + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_s2g", "compatibility_paths": [ - "cuda_device::tma::cp_async_bulk_tensor_2d_s2g" + "cuda_device::tma::cp_async_bulk_tensor_3d_s2g" ] }, "dialect": { - "op_type": "CpAsyncBulkTensorS2gTile2dOp", - "op_name": "nvvm.cp_async_bulk_tensor_s2g_tile_2d", + "op_type": "CpAsyncBulkTensorS2gTile3dOp", + "op_name": "nvvm.cp_async_bulk_tensor_s2g_tile_3d", "operands": [ "ptr", "ptr", "i32", + "i32", "i32" ], "results": [] }, "llvm": { - "symbol": "llvm.nvvm.cp.async.bulk.tensor.s2g.tile.2d", + "symbol": "llvm.nvvm.cp.async.bulk.tensor.s2g.tile.3d", "arguments": [ "shared_ptr", "ptr", "i32", "i32", + "i32", "i64", "i1" ], "results": [], "properties": [ - "ImmArg", + "ImmArg", "IntrConvergent", "ReadOnly", "ReadOnly" @@ -19940,7 +21307,7 @@ } ], "tma": { - "operation": "s2g_tile2d", + "operation": "s2g_tile3d", "adapter": "s2g_pointers_coordinates_inject_defaults", "runtime_validation": "unexecuted" }, @@ -19951,7 +21318,7 @@ "async", "bulk", "tensor", - "2d", + "3d", "global", "shared::cta", "tile", @@ -19969,17 +21336,17 @@ "summary": "Starts a TMA tile copy from shared to global memory." }, { - "id": "cp_async_bulk_tensor_3d_g2s", - "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.3d", + "id": "cp_async_bulk_tensor_4d_g2s", + "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.4d", "family": "tma", "source": { "kind": "llvm_imported", - "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_3d" + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_4d" }, "selections": [ { - "source_record": "TMA_G2S_TILE_CG0_3D", - "asm": "cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2}}], [$mbar];", + "source_record": "TMA_G2S_TILE_CG0_4D", + "asm": "cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2, $d3}}], [$mbar];", "predicates": [ "Subtarget->getPTXVersion() >= 80", "Subtarget->getSmVersion() >= 90" @@ -19987,11 +21354,11 @@ "constraints": { "immediate_bindings": [ { - "argument_index": 8, + "argument_index": 9, "value": 0 }, { - "argument_index": 9, + "argument_index": 10, "value": 0 } ] @@ -19999,30 +21366,31 @@ } ], "rust": { - "abi_id": "i0332", + "abi_id": "i0333", "module": "tma", - "name": "cp_async_bulk_tensor_3d_g2s", + "name": "cp_async_bulk_tensor_4d_g2s", "arguments": [ "*mut u8", "*const u8", "i32", "i32", "i32", + "i32", "*mut u64" ], "result": "()", "safe": false, "must_use": false, "safe_allowlist_reason": null, - "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0332", - "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s", + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0333", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s", "compatibility_paths": [ - "cuda_device::tma::cp_async_bulk_tensor_3d_g2s" + "cuda_device::tma::cp_async_bulk_tensor_4d_g2s" ] }, "dialect": { - "op_type": "CpAsyncBulkTensorG2sTile3dOp", - "op_name": "nvvm.cp_async_bulk_tensor_g2s_tile_3d", + "op_type": "CpAsyncBulkTensorG2sTile4dOp", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_tile_4d", "operands": [ "ptr", "ptr", @@ -20030,13 +21398,14 @@ "i32", "i32", "i32", + "i32", "i16", "i64" ], "results": [] }, "llvm": { - "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.3d", + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.4d", "arguments": [ "shared_cluster_ptr", "shared_ptr", @@ -20044,6 +21413,7 @@ "i32", "i32", "i32", + "i32", "i16", "i64", "i1", @@ -20053,10 +21423,10 @@ "results": [], "properties": [ "ImmArg", - "ImmArg", + "ImmArg", "ImmArg", "IntrConvergent", - "Range", + "Range", "ReadOnly", "WriteOnly" ], @@ -20222,7 +21592,7 @@ } ], "tma": { - "operation": "g2s_tile3d", + "operation": "g2s_tile4d", "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", "runtime_validation": "unexecuted" }, @@ -20233,7 +21603,7 @@ "async", "bulk", "tensor", - "3d", + "4d", "shared::cluster", "global", "tile", @@ -20254,25 +21624,25 @@ "summary": "Starts a TMA tile copy from global to cluster shared memory." }, { - "id": "cp_async_bulk_tensor_3d_s2g", - "operation_key": "memory.copy.async.bulk.tensor.s2g.tile.3d", + "id": "cp_async_bulk_tensor_4d_g2s_cta", + "operation_key": "memory.copy.async.bulk.tensor.g2s.cta.tile.4d", "family": "tma", "source": { "kind": "llvm_imported", - "source_record": "int_nvvm_cp_async_bulk_tensor_s2g_tile_3d" + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_4d" }, "selections": [ { - "source_record": "TMA_TENSOR_S2G_TILE_3D", - "asm": "cp.async.bulk.tensor.3d.global.shared::cta.tile.bulk_group [$tmap, {{$d0, $d1, $d2}}], [$src];", + "source_record": "TMA_G2S_CTA_TILE_4D", + "asm": "cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2, $d3}}], [$mbar];", "predicates": [ - "Subtarget->getPTXVersion() >= 80", - "Subtarget->getSmVersion() >= 90" + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" ], "constraints": { "immediate_bindings": [ { - "argument_index": 6, + "argument_index": 8, "value": 0 } ] @@ -20280,55 +21650,62 @@ } ], "rust": { - "abi_id": "i0337", + "abi_id": "i1037", "module": "tma", - "name": "cp_async_bulk_tensor_3d_s2g", + "name": "cp_async_bulk_tensor_4d_g2s_cta", "arguments": [ - "*const u8", + "*mut u8", "*const u8", "i32", "i32", - "i32" + "i32", + "i32", + "*mut u64" ], "result": "()", "safe": false, "must_use": false, "safe_allowlist_reason": null, - "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337", - "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_3d_s2g", + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s_cta", "compatibility_paths": [ - "cuda_device::tma::cp_async_bulk_tensor_3d_s2g" + "cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta" ] }, "dialect": { - "op_type": "CpAsyncBulkTensorS2gTile3dOp", - "op_name": "nvvm.cp_async_bulk_tensor_s2g_tile_3d", + "op_type": "CpAsyncBulkTensorG2sCtaTile4dOp", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_cta_tile_4d", "operands": [ + "ptr", "ptr", "ptr", "i32", "i32", - "i32" + "i32", + "i32", + "i64" ], "results": [] }, "llvm": { - "symbol": "llvm.nvvm.cp.async.bulk.tensor.s2g.tile.3d", + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.4d", "arguments": [ + "shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", + "i32", "i64", "i1" ], "results": [], "properties": [ - "ImmArg", + "ImmArg", "IntrConvergent", - "ReadOnly", - "ReadOnly" + "ReadOnly", + "WriteOnly" ], "result_facts": { "no_undef": false, @@ -20342,7 +21719,7 @@ "execution_scope": "thread" }, "target": { - "minimum_ptx": "8.0", + "minimum_ptx": "8.6", "hardware": { "kind": "any_of", "alternatives": [ @@ -20359,21 +21736,21 @@ "ptx_isa_url": "https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor" }, "backend": { - "profile": "rust-llvm-23.1.3-aab11ee0-tma", + "profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", "version": "LLVM version 23.1.3-rust-1.101.0-nightly", "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", "status": "validated", "target_triple": "nvptx64-nvidia-cuda", "gpu_target": "sm_90", - "ptx_feature": "+ptx80" + "ptx_feature": "+ptx86" }, "backend_lowerings": [ { "backend": "llvm_nvptx", "mechanism": "typed_nvvm", - "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma", + "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", "target": { - "minimum_ptx": "8.0", + "minimum_ptx": "8.6", "hardware": { "kind": "any_of", "alternatives": [ @@ -20394,11 +21771,11 @@ "targets": [ "sm_90" ], - "representation": "separate typed LLVM declaration probes", + "representation": "separate typed LLVM declaration probe for the 4D CTA-local TMA G2S operation", "stage": "declaration_canonicalization", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581.", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_4d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,100-byte source had SHA-256 becebd363b47d76e388fccf46646db4d8bba5225b4eb23cc7cae1673a5411ddd; the 2,664-byte bitcode had SHA-256 77f23ed44102d830fb8a576a8ff62f947e96b2b15871e2984083851d1e82f63e; the 1,327-byte canonical IR had SHA-256 6fd6a38abbb214d1e1fc5a4c8aa800ab963d1f53350afc3a7ec965501e467fb4.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -20407,13 +21784,13 @@ { "targets": [ "sm_90", - "ptx80" + "ptx86" ], - "representation": "separate probes for the recorded lowering mechanism", + "representation": "typed LLVM call for the 4D CTA-local TMA G2S operation", "stage": "backend_codegen", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): the exact reviewed PTX instruction shapes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 418eaf16785cbdd0e5b8fb5d886e2768f4347b8dda669cd924ad0cd857e7e08a.", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,567-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 0607dfca39829f02dbda9367602cb27d28e6dc5dd8f5909a451fa4648442629a.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -20422,13 +21799,13 @@ { "targets": [ "sm_90", - "ptx80" + "ptx86" ], - "representation": "separate PTX outputs assembled for the recorded target", + "representation": "LLVM-produced PTX retaining the 4D CTA-local TMA G2S instruction", "stage": "ptx_assembly", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): terminal PTX assembly passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 45075184f52be336a8f24e8232251aec1a5890d9a6f0d35cbef2602dca1dafaa.", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", "artifact_kind": "cubin", "tool_path": "/usr/local/cuda-13.3/bin/ptxas", "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", @@ -20439,9 +21816,9 @@ { "backend": "lib_nvvm", "mechanism": "inline_ptx", - "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma", + "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma-cta", "target": { - "minimum_ptx": "8.0", + "minimum_ptx": "8.6", "hardware": { "kind": "any_of", "alternatives": [ @@ -20463,11 +21840,11 @@ "compute_90", "ptx93" ], - "representation": "14 per-operation NVVM IR 2.0 modules covering the non-CTA-pair inline-PTX forms", + "representation": "one NVVM IR 2.0 module containing the five CTA-local TMA G2S inline-PTX forms", "stage": "backend_codegen", "mechanism": "inline_ptx", "outcome": "succeeded", - "detail": "CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX.", + "detail": "CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -20478,11 +21855,11 @@ "sm_90", "ptx93" ], - "representation": "14 linked cubins covering the non-CTA-pair TMA instructions", + "representation": "linked cubin retaining five UTMALDG instructions", "stage": "device_link", "mechanism": "inline_ptx", "outcome": "succeeded", - "detail": "CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303.", + "detail": "CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions.", "artifact_kind": "cubin", "tool_path": "libnvJitLink.so.13.3.33", "tool_version": "CUDA 13.3 nvJitLink 13.3.33", @@ -20492,8 +21869,8 @@ } ], "tma": { - "operation": "s2g_tile3d", - "adapter": "s2g_pointers_coordinates_inject_defaults", + "operation": "g2s_cta_tile4d", + "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", "runtime_validation": "unexecuted" }, "lowering": "generated_tma", @@ -20503,35 +21880,38 @@ "async", "bulk", "tensor", - "3d", - "global", + "4d", "shared::cta", + "global", "tile", - "bulk_group" + "mbarrier::complete_tx::bytes" ], "operands": [ { "kind": "address" }, + { + "kind": "address" + }, { "kind": "address" } ] }, - "summary": "Starts a TMA tile copy from shared to global memory." + "summary": "Starts a TMA tile copy from global to the issuing CTA's shared memory." }, { - "id": "cp_async_bulk_tensor_4d_g2s", - "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.4d", + "id": "cp_async_bulk_tensor_4d_s2g", + "operation_key": "memory.copy.async.bulk.tensor.s2g.tile.4d", "family": "tma", "source": { "kind": "llvm_imported", - "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_4d" + "source_record": "int_nvvm_cp_async_bulk_tensor_s2g_tile_4d" }, "selections": [ { - "source_record": "TMA_G2S_TILE_CG0_4D", - "asm": "cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2, $d3}}], [$mbar];", + "source_record": "TMA_TENSOR_S2G_TILE_4D", + "asm": "cp.async.bulk.tensor.4d.global.shared::cta.tile.bulk_group [$tmap, {{$d0, $d1, $d2, $d3}}], [$src];", "predicates": [ "Subtarget->getPTXVersion() >= 80", "Subtarget->getSmVersion() >= 90" @@ -20539,11 +21919,7 @@ "constraints": { "immediate_bindings": [ { - "argument_index": 9, - "value": 0 - }, - { - "argument_index": 10, + "argument_index": 7, "value": 0 } ] @@ -20551,69 +21927,58 @@ } ], "rust": { - "abi_id": "i0333", + "abi_id": "i0338", "module": "tma", - "name": "cp_async_bulk_tensor_4d_g2s", + "name": "cp_async_bulk_tensor_4d_s2g", "arguments": [ - "*mut u8", + "*const u8", "*const u8", "i32", "i32", "i32", - "i32", - "*mut u64" + "i32" ], "result": "()", "safe": false, "must_use": false, "safe_allowlist_reason": null, - "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0333", - "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s", + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_s2g", "compatibility_paths": [ - "cuda_device::tma::cp_async_bulk_tensor_4d_g2s" + "cuda_device::tma::cp_async_bulk_tensor_4d_s2g" ] }, "dialect": { - "op_type": "CpAsyncBulkTensorG2sTile4dOp", - "op_name": "nvvm.cp_async_bulk_tensor_g2s_tile_4d", + "op_type": "CpAsyncBulkTensorS2gTile4dOp", + "op_name": "nvvm.cp_async_bulk_tensor_s2g_tile_4d", "operands": [ - "ptr", "ptr", "ptr", "i32", "i32", "i32", - "i32", - "i16", - "i64" + "i32" ], "results": [] }, "llvm": { - "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.4d", + "symbol": "llvm.nvvm.cp.async.bulk.tensor.s2g.tile.4d", "arguments": [ - "shared_cluster_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", "i32", - "i16", "i64", - "i1", - "i1", - "i32" + "i1" ], "results": [], "properties": [ - "ImmArg", - "ImmArg", - "ImmArg", + "ImmArg", "IntrConvergent", - "Range", - "ReadOnly", - "WriteOnly" + "ReadOnly", + "ReadOnly" ], "result_facts": { "no_undef": false, @@ -20777,8 +22142,8 @@ } ], "tma": { - "operation": "g2s_tile4d", - "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", + "operation": "s2g_tile4d", + "adapter": "s2g_pointers_coordinates_inject_defaults", "runtime_validation": "unexecuted" }, "lowering": "generated_tma", @@ -20789,37 +22154,34 @@ "bulk", "tensor", "4d", - "shared::cluster", "global", + "shared::cta", "tile", - "mbarrier::complete_tx::bytes" + "bulk_group" ], "operands": [ { "kind": "address" }, - { - "kind": "address" - }, { "kind": "address" } ] }, - "summary": "Starts a TMA tile copy from global to cluster shared memory." + "summary": "Starts a TMA tile copy from shared to global memory." }, { - "id": "cp_async_bulk_tensor_4d_s2g", - "operation_key": "memory.copy.async.bulk.tensor.s2g.tile.4d", + "id": "cp_async_bulk_tensor_5d_g2s", + "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.5d", "family": "tma", "source": { "kind": "llvm_imported", - "source_record": "int_nvvm_cp_async_bulk_tensor_s2g_tile_4d" + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_5d" }, "selections": [ { - "source_record": "TMA_TENSOR_S2G_TILE_4D", - "asm": "cp.async.bulk.tensor.4d.global.shared::cta.tile.bulk_group [$tmap, {{$d0, $d1, $d2, $d3}}], [$src];", + "source_record": "TMA_G2S_TILE_CG0_5D", + "asm": "cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2, $d3, $d4}}], [$mbar];", "predicates": [ "Subtarget->getPTXVersion() >= 80", "Subtarget->getSmVersion() >= 90" @@ -20827,7 +22189,11 @@ "constraints": { "immediate_bindings": [ { - "argument_index": 7, + "argument_index": 10, + "value": 0 + }, + { + "argument_index": 11, "value": 0 } ] @@ -20835,58 +22201,72 @@ } ], "rust": { - "abi_id": "i0338", + "abi_id": "i0334", "module": "tma", - "name": "cp_async_bulk_tensor_4d_s2g", + "name": "cp_async_bulk_tensor_5d_g2s", "arguments": [ - "*const u8", + "*mut u8", "*const u8", "i32", "i32", "i32", - "i32" + "i32", + "i32", + "*mut u64" ], "result": "()", "safe": false, "must_use": false, "safe_allowlist_reason": null, - "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338", - "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_4d_s2g", + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0334", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s", "compatibility_paths": [ - "cuda_device::tma::cp_async_bulk_tensor_4d_s2g" + "cuda_device::tma::cp_async_bulk_tensor_5d_g2s" ] }, "dialect": { - "op_type": "CpAsyncBulkTensorS2gTile4dOp", - "op_name": "nvvm.cp_async_bulk_tensor_s2g_tile_4d", + "op_type": "CpAsyncBulkTensorG2sTile5dOp", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_tile_5d", "operands": [ + "ptr", "ptr", "ptr", "i32", "i32", "i32", - "i32" + "i32", + "i32", + "i16", + "i64" ], "results": [] }, "llvm": { - "symbol": "llvm.nvvm.cp.async.bulk.tensor.s2g.tile.4d", + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.5d", "arguments": [ + "shared_cluster_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", "i32", + "i32", + "i16", "i64", - "i1" + "i1", + "i1", + "i32" ], "results": [], "properties": [ - "ImmArg", + "ImmArg", + "ImmArg", + "ImmArg", "IntrConvergent", - "ReadOnly", - "ReadOnly" + "Range", + "ReadOnly", + "WriteOnly" ], "result_facts": { "no_undef": false, @@ -21050,8 +22430,8 @@ } ], "tma": { - "operation": "s2g_tile4d", - "adapter": "s2g_pointers_coordinates_inject_defaults", + "operation": "g2s_tile5d", + "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", "runtime_validation": "unexecuted" }, "lowering": "generated_tma", @@ -21061,47 +22441,46 @@ "async", "bulk", "tensor", - "4d", + "5d", + "shared::cluster", "global", - "shared::cta", "tile", - "bulk_group" + "mbarrier::complete_tx::bytes" ], "operands": [ { "kind": "address" }, + { + "kind": "address" + }, { "kind": "address" } ] }, - "summary": "Starts a TMA tile copy from shared to global memory." + "summary": "Starts a TMA tile copy from global to cluster shared memory." }, { - "id": "cp_async_bulk_tensor_5d_g2s", - "operation_key": "memory.copy.async.bulk.tensor.g2s.tile.5d", + "id": "cp_async_bulk_tensor_5d_g2s_cta", + "operation_key": "memory.copy.async.bulk.tensor.g2s.cta.tile.5d", "family": "tma", "source": { "kind": "llvm_imported", - "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_tile_5d" + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_5d" }, "selections": [ { - "source_record": "TMA_G2S_TILE_CG0_5D", - "asm": "cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes$cg [$dst], [$tmap, {{$d0, $d1, $d2, $d3, $d4}}], [$mbar];", + "source_record": "TMA_G2S_CTA_TILE_5D", + "asm": "cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes [$dst], [$tmap, {{$d0, $d1, $d2, $d3, $d4}}], [$mbar];", "predicates": [ - "Subtarget->getPTXVersion() >= 80", - "Subtarget->getSmVersion() >= 90" + "Subtarget->getSmVersion() >= 90", + "Subtarget->getPTXVersion() >= 86" ], "constraints": { "immediate_bindings": [ { - "argument_index": 10, - "value": 0 - }, - { - "argument_index": 11, + "argument_index": 9, "value": 0 } ] @@ -21109,9 +22488,9 @@ } ], "rust": { - "abi_id": "i0334", + "abi_id": "i1038", "module": "tma", - "name": "cp_async_bulk_tensor_5d_g2s", + "name": "cp_async_bulk_tensor_5d_g2s_cta", "arguments": [ "*mut u8", "*const u8", @@ -21126,15 +22505,15 @@ "safe": false, "must_use": false, "safe_allowlist_reason": null, - "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0334", - "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s", + "canonical_path": "cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038", + "public_path": "cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s_cta", "compatibility_paths": [ - "cuda_device::tma::cp_async_bulk_tensor_5d_g2s" + "cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta" ] }, "dialect": { - "op_type": "CpAsyncBulkTensorG2sTile5dOp", - "op_name": "nvvm.cp_async_bulk_tensor_g2s_tile_5d", + "op_type": "CpAsyncBulkTensorG2sCtaTile5dOp", + "op_name": "nvvm.cp_async_bulk_tensor_g2s_cta_tile_5d", "operands": [ "ptr", "ptr", @@ -21144,15 +22523,14 @@ "i32", "i32", "i32", - "i16", "i64" ], "results": [] }, "llvm": { - "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.tile.5d", + "symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.5d", "arguments": [ - "shared_cluster_ptr", + "shared_ptr", "shared_ptr", "ptr", "i32", @@ -21160,19 +22538,13 @@ "i32", "i32", "i32", - "i16", "i64", - "i1", - "i1", - "i32" + "i1" ], "results": [], "properties": [ - "ImmArg", - "ImmArg", - "ImmArg", + "ImmArg", "IntrConvergent", - "Range", "ReadOnly", "WriteOnly" ], @@ -21188,7 +22560,7 @@ "execution_scope": "thread" }, "target": { - "minimum_ptx": "8.0", + "minimum_ptx": "8.6", "hardware": { "kind": "any_of", "alternatives": [ @@ -21205,21 +22577,21 @@ "ptx_isa_url": "https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor" }, "backend": { - "profile": "rust-llvm-23.1.3-aab11ee0-tma", + "profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", "version": "LLVM version 23.1.3-rust-1.101.0-nightly", "sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", "status": "validated", "target_triple": "nvptx64-nvidia-cuda", "gpu_target": "sm_90", - "ptx_feature": "+ptx80" + "ptx_feature": "+ptx86" }, "backend_lowerings": [ { "backend": "llvm_nvptx", "mechanism": "typed_nvvm", - "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma", + "evidence_profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", "target": { - "minimum_ptx": "8.0", + "minimum_ptx": "8.6", "hardware": { "kind": "any_of", "alternatives": [ @@ -21240,11 +22612,11 @@ "targets": [ "sm_90" ], - "representation": "separate typed LLVM declaration probes", + "representation": "separate typed LLVM declaration probe for the 5D CTA-local TMA G2S operation", "stage": "declaration_canonicalization", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581.", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_5d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,131-byte source had SHA-256 da8bf49e04c4250626e97150345c704f33802066582d2e80d1c17dd58c0fa0b0; the 2,668-byte bitcode had SHA-256 d1d535b89cc23b4983c8fb5b00562e06266454177d8eb3a81684859ac5430cb5; the 1,358-byte canonical IR had SHA-256 e4862fd8477461d79f65871ae6dc8750d273a0e50911602e0bfea91bae76df7c.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -21253,13 +22625,13 @@ { "targets": [ "sm_90", - "ptx80" + "ptx86" ], - "representation": "separate probes for the recorded lowering mechanism", + "representation": "typed LLVM call for the 5D CTA-local TMA G2S operation", "stage": "backend_codegen", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): the exact reviewed PTX instruction shapes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 418eaf16785cbdd0e5b8fb5d886e2768f4347b8dda669cd924ad0cd857e7e08a.", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,701-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 2f87ab7dc1ed68335e4ad29e4bcc7bd264779b33deb26f1d18e0c955f6bf8223.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -21268,13 +22640,13 @@ { "targets": [ "sm_90", - "ptx80" + "ptx86" ], - "representation": "separate PTX outputs assembled for the recorded target", + "representation": "LLVM-produced PTX retaining the 5D CTA-local TMA G2S instruction", "stage": "ptx_assembly", "mechanism": "typed_nvvm", "outcome": "succeeded", - "detail": "Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): terminal PTX assembly passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 45075184f52be336a8f24e8232251aec1a5890d9a6f0d35cbef2602dca1dafaa.", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", "artifact_kind": "cubin", "tool_path": "/usr/local/cuda-13.3/bin/ptxas", "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", @@ -21285,9 +22657,9 @@ { "backend": "lib_nvvm", "mechanism": "inline_ptx", - "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma", + "evidence_profile": "cuda-13.3-libnvvm-13.3.73-tma-cta", "target": { - "minimum_ptx": "8.0", + "minimum_ptx": "8.6", "hardware": { "kind": "any_of", "alternatives": [ @@ -21309,11 +22681,11 @@ "compute_90", "ptx93" ], - "representation": "14 per-operation NVVM IR 2.0 modules covering the non-CTA-pair inline-PTX forms", + "representation": "one NVVM IR 2.0 module containing the five CTA-local TMA G2S inline-PTX forms", "stage": "backend_codegen", "mechanism": "inline_ptx", "outcome": "succeeded", - "detail": "CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX.", + "detail": "CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60.", "artifact_kind": null, "tool_path": null, "tool_version": null, @@ -21324,11 +22696,11 @@ "sm_90", "ptx93" ], - "representation": "14 linked cubins covering the non-CTA-pair TMA instructions", + "representation": "linked cubin retaining five UTMALDG instructions", "stage": "device_link", "mechanism": "inline_ptx", "outcome": "succeeded", - "detail": "CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303.", + "detail": "CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions.", "artifact_kind": "cubin", "tool_path": "libnvJitLink.so.13.3.33", "tool_version": "CUDA 13.3 nvJitLink 13.3.33", @@ -21338,7 +22710,7 @@ } ], "tma": { - "operation": "g2s_tile5d", + "operation": "g2s_cta_tile5d", "adapter": "g2s_pointers_coordinates_barrier_inject_defaults", "runtime_validation": "unexecuted" }, @@ -21350,7 +22722,7 @@ "bulk", "tensor", "5d", - "shared::cluster", + "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes" @@ -21367,7 +22739,7 @@ } ] }, - "summary": "Starts a TMA tile copy from global to cluster shared memory." + "summary": "Starts a TMA tile copy from global to the issuing CTA's shared memory." }, { "id": "cp_async_bulk_tensor_5d_s2g", diff --git a/cuda-oxide/intrinsics/evidence/cuda-13.3-libnvvm-13.3.73-tma-cta.json b/cuda-oxide/intrinsics/evidence/cuda-13.3-libnvvm-13.3.73-tma-cta.json new file mode 100644 index 000000000..21d38463e --- /dev/null +++ b/cuda-oxide/intrinsics/evidence/cuda-13.3-libnvvm-13.3.73-tma-cta.json @@ -0,0 +1,109 @@ +{ + "schema": 6, + "backend_profile": "cuda-13.3-libnvvm-13.3.73-tma-cta", + "backend_kind": "lib_nvvm", + "llvm_revision": "cuda-13.3-libnvvm", + "backend_version": "CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0", + "backend_sha256": "5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc", + "build_id_prefix": "88b802", + "nvvm_ir_version": "2.0", + "debug_ir_version": "3.2", + "defaults": { + "llvm_results": [], + "target_triple": "nvptx64-nvidia-cuda", + "gpu_target": "compute_90", + "ptx_feature": "PTX ISA 9.3 emitted by CUDA 13.3", + "status": "validated", + "runtime_validation": "unexecuted" + }, + "fixtures": [ + { + "id": "tma_cta_inline_sm90", + "coverage_count": 5, + "stages": [ + { + "targets": ["compute_90", "ptx93"], + "representation": "one NVVM IR 2.0 module containing the five CTA-local TMA G2S inline-PTX forms", + "stage": "backend_codegen", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "detail": "CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60." + }, + { + "targets": ["sm_90", "ptx93"], + "representation": "linked cubin retaining five UTMALDG instructions", + "stage": "device_link", + "mechanism": "inline_ptx", + "outcome": "succeeded", + "artifact_kind": "cubin", + "detail": "CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions.", + "tool_path": "libnvJitLink.so.13.3.33", + "tool_version": "CUDA 13.3 nvJitLink 13.3.33", + "tool_sha256": "f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa" + } + ] + } + ], + "matrices": [ + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_1d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_inline_sm90"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_1d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.1d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "1d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i64", "i1"]} + } + }, + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_2d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_inline_sm90"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_2d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "2d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i64", "i1"]} + } + }, + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_3d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_inline_sm90"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_3d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.3d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "3d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", "i64", "i1"]} + } + }, + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_4d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_inline_sm90"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_4d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.4d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "4d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", "i32", "i64", "i1"]} + } + }, + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_5d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_inline_sm90"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_5d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.5d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "5d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", "i32", "i32", "i64", "i1"]} + } + } + ] +} diff --git a/cuda-oxide/intrinsics/evidence/rust-llvm-23.1.3-aab11ee0-tma-cta.json b/cuda-oxide/intrinsics/evidence/rust-llvm-23.1.3-aab11ee0-tma-cta.json new file mode 100644 index 000000000..a4ea137c6 --- /dev/null +++ b/cuda-oxide/intrinsics/evidence/rust-llvm-23.1.3-aab11ee0-tma-cta.json @@ -0,0 +1,252 @@ +{ + "schema": 6, + "backend_profile": "rust-llvm-23.1.3-aab11ee0-tma-cta", + "backend_kind": "llvm_nvptx", + "llvm_revision": "aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019", + "backend_version": "LLVM version 23.1.3-rust-1.101.0-nightly", + "backend_sha256": "7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449", + "defaults": { + "llvm_results": [], + "concrete_llvm_results": [], + "target_triple": "nvptx64-nvidia-cuda", + "gpu_target": "sm_90", + "ptx_feature": "+ptx86", + "status": "validated", + "declaration_attributes_canonicalized": true, + "runtime_validation": "unexecuted" + }, + "fixtures": [ + { + "id": "tma_cta_typed_sm90_1d", + "coverage_count": 1, + "stages": [ + { + "targets": ["sm_90"], + "representation": "separate typed LLVM declaration probe for the 1D CTA-local TMA G2S operation", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_1d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,007-byte source had SHA-256 253d3ea39f9d26f5000f83febb9f72a15af2bed6ad0f668ab206976ce4c6d6ba; the 2,640-byte bitcode had SHA-256 9f42944557579c5c9c4e38716cce52607a07c51e5f2272393948fdc40cfb8961; the 1,234-byte canonical IR had SHA-256 835c2efc3b53d5333ec6a8260fcad29d5d0ce889194d0523feb57f7b22281fff." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "typed LLVM call for the 1D CTA-local TMA G2S operation", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,165-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 b0b4469f60fea3342f58e26eb4698834c6cb0714aac17fd3ffaed79cbd4f226a." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "LLVM-produced PTX retaining the 1D CTA-local TMA G2S instruction", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "artifact_kind": "cubin", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "id": "tma_cta_typed_sm90_2d", + "coverage_count": 1, + "stages": [ + { + "targets": ["sm_90"], + "representation": "separate typed LLVM declaration probe for the 2D CTA-local TMA G2S operation", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_2d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,038-byte source had SHA-256 f0d47c1b6670854dfddce5bc45206d8f4ad16e9a05236dd4e1d4036b518666ea; the 2,648-byte bitcode had SHA-256 6ab53bfd2185eb8b6d7f7ffb3793d72cb7b48a07f61846a5317604700bdd3225; the 1,265-byte canonical IR had SHA-256 2c00613205683c093c5002f0f7347d3b2d6b82e46c5bc00072fb5a781f7d9d44." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "typed LLVM call for the 2D CTA-local TMA G2S operation", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,299-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 59fa9746f48c669cbff59faffea6039a796055f2721b8647a6331d78b5a26521." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "LLVM-produced PTX retaining the 2D CTA-local TMA G2S instruction", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "artifact_kind": "cubin", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "id": "tma_cta_typed_sm90_3d", + "coverage_count": 1, + "stages": [ + { + "targets": ["sm_90"], + "representation": "separate typed LLVM declaration probe for the 3D CTA-local TMA G2S operation", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_3d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,069-byte source had SHA-256 159ab50347ad4f81c8a110df7de35db35da9eb099938874523e5ec10ea867c8c; the 2,656-byte bitcode had SHA-256 3b5f19c93cffadf279efad33b501519f09cacbc660784f195ab3d007b5bca854; the 1,296-byte canonical IR had SHA-256 223dd701edae4ecf145ca18afc659dd5a50785cf8e492f6ba5c7b7f8299dc030." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "typed LLVM call for the 3D CTA-local TMA G2S operation", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,433-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 e9bab5f8e2c2a697ca435f86eaff468984a03470ae4bf28f975c2c89fc6676b9." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "LLVM-produced PTX retaining the 3D CTA-local TMA G2S instruction", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "artifact_kind": "cubin", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "id": "tma_cta_typed_sm90_4d", + "coverage_count": 1, + "stages": [ + { + "targets": ["sm_90"], + "representation": "separate typed LLVM declaration probe for the 4D CTA-local TMA G2S operation", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_4d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,100-byte source had SHA-256 becebd363b47d76e388fccf46646db4d8bba5225b4eb23cc7cae1673a5411ddd; the 2,664-byte bitcode had SHA-256 77f23ed44102d830fb8a576a8ff62f947e96b2b15871e2984083851d1e82f63e; the 1,327-byte canonical IR had SHA-256 6fd6a38abbb214d1e1fc5a4c8aa800ab963d1f53350afc3a7ec965501e467fb4." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "typed LLVM call for the 4D CTA-local TMA G2S operation", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,567-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 0607dfca39829f02dbda9367602cb27d28e6dc5dd8f5909a451fa4648442629a." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "LLVM-produced PTX retaining the 4D CTA-local TMA G2S instruction", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "artifact_kind": "cubin", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + }, + { + "id": "tma_cta_typed_sm90_5d", + "coverage_count": 1, + "stages": [ + { + "targets": ["sm_90"], + "representation": "separate typed LLVM declaration probe for the 5D CTA-local TMA G2S operation", + "stage": "declaration_canonicalization", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_5d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,131-byte source had SHA-256 da8bf49e04c4250626e97150345c704f33802066582d2e80d1c17dd58c0fa0b0; the 2,668-byte bitcode had SHA-256 d1d535b89cc23b4983c8fb5b00562e06266454177d8eb3a81684859ac5430cb5; the 1,358-byte canonical IR had SHA-256 e4862fd8477461d79f65871ae6dc8750d273a0e50911602e0bfea91bae76df7c." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "typed LLVM call for the 5D CTA-local TMA G2S operation", + "stage": "backend_codegen", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "detail": "The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,701-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 2f87ab7dc1ed68335e4ad29e4bcc7bd264779b33deb26f1d18e0c955f6bf8223." + }, + { + "targets": ["sm_90", "ptx86"], + "representation": "LLVM-produced PTX retaining the 5D CTA-local TMA G2S instruction", + "stage": "ptx_assembly", + "mechanism": "typed_nvvm", + "outcome": "succeeded", + "artifact_kind": "cubin", + "detail": "CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480.", + "tool_path": "/usr/local/cuda-13.3/bin/ptxas", + "tool_version": "Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0", + "tool_sha256": "afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc" + } + ] + } + ], + "matrices": [ + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_1d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_typed_sm90_1d"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_1d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.1d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "1d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i64", "i1"], "concrete_llvm_arguments": ["ptr addrspace(3)", "ptr addrspace(3)", "ptr", "i32", "i64", "i1"]} + } + }, + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_2d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_typed_sm90_2d"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_2d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "2d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i64", "i1"], "concrete_llvm_arguments": ["ptr addrspace(3)", "ptr addrspace(3)", "ptr", "i32", "i32", "i64", "i1"]} + } + }, + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_3d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_typed_sm90_3d"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_3d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.3d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "3d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", "i64", "i1"], "concrete_llvm_arguments": ["ptr addrspace(3)", "ptr addrspace(3)", "ptr", "i32", "i32", "i32", "i64", "i1"]} + } + }, + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_4d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_typed_sm90_4d"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_4d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.4d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "4d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", "i32", "i64", "i1"], "concrete_llvm_arguments": ["ptr addrspace(3)", "ptr addrspace(3)", "ptr", "i32", "i32", "i32", "i32", "i64", "i1"]} + } + }, + { + "axes": [{"name": "id", "values": ["cp_async_bulk_tensor_5d_g2s_cta"]}], + "product_count": 1, + "fixtures": ["tma_cta_typed_sm90_5d"], + "template": { + "id": "${id}", + "source_record": "int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_5d", + "llvm_symbol": "llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.5d", + "expected_ptx": {"mnemonic": "cp", "modifiers": ["async", "bulk", "tensor", "5d", "shared::cta", "global", "tile", "mbarrier::complete_tx::bytes"], "operands": [{"kind": "address"}, {"kind": "address"}, {"kind": "address"}]}, + "facts": {"llvm_arguments": ["shared_ptr", "shared_ptr", "ptr", "i32", "i32", "i32", "i32", "i32", "i64", "i1"], "concrete_llvm_arguments": ["ptr addrspace(3)", "ptr addrspace(3)", "ptr", "i32", "i32", "i32", "i32", "i32", "i64", "i1"]} + } + } + ] +} diff --git a/cuda-oxide/intrinsics/generated-reference.md b/cuda-oxide/intrinsics/generated-reference.md index fbcbe1298..444c16376 100644 --- a/cuda-oxide/intrinsics/generated-reference.md +++ b/cuda-oxide/intrinsics/generated-reference.md @@ -2,7 +2,7 @@ SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. SPDX-License-Identifier: Apache-2.0 @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. --> @@ -97,16 +97,21 @@ This table is generated from the resolved catalog. Evidence stages below disting | `cuda_intrinsics::tma::cp_async_bulk_prefetch_tensor_gather4_2d_l2` | `nvvm.cp_async_bulk_prefetch_tensor_gather4_2d_l2` | LLVM `llvm.nvvm.cp.async.bulk.tensor.prefetch.tile.gather4.2d` from `int_nvvm_cp_async_bulk_tensor_prefetch_tile_gather4_2d` | unsafe; scope thread; impure; memory read; convergent true | PTX 8.6 on sm_100a or sm_101a or sm_103a or sm_110a ([PTX ISA 9.3, Tensor-map descriptor and asynchronous bulk tensor operations](https://docs.nvidia.com/cuda/parallel-thread-execution/#tensor-map)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.prefetch.tensor.2d.L2.global.tile::gather4 [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint` | `nvvm.cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint` | LLVM `llvm.nvvm.cp.async.bulk.tensor.prefetch.tile.gather4.2d` from `int_nvvm_cp_async_bulk_tensor_prefetch_tile_gather4_2d` | unsafe; scope thread; impure; memory read; convergent true | PTX 8.6 on sm_100a or sm_101a or sm_103a or sm_110a ([PTX ISA 9.3, Tensor-map descriptor and asynchronous bulk tensor operations](https://docs.nvidia.com/cuda/parallel-thread-execution/#tensor-map)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.prefetch.tensor.2d.L2.global.tile::gather4.L2::cache_hint [
], ;` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_1d_g2s` | `nvvm.cp_async_bulk_tensor_g2s_tile_1d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.tile.1d` from `int_nvvm_cp_async_bulk_tensor_g2s_tile_1d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | +| `cuda_intrinsics::tma::cp_async_bulk_tensor_1d_g2s_cta` | `nvvm.cp_async_bulk_tensor_g2s_cta_tile_1d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.1d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_1d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.6 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma-cta`; expects `cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_1d_s2g` | `nvvm.cp_async_bulk_tensor_s2g_tile_1d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.1d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_1d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.1d.global.shared::cta.tile.bulk_group [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s` | `nvvm.cp_async_bulk_tensor_g2s_tile_2d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d` from `int_nvvm_cp_async_bulk_tensor_g2s_tile_2d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | +| `cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_cta` | `nvvm.cp_async_bulk_tensor_g2s_cta_tile_2d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_2d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.6 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma-cta`; expects `cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_multicast` | `nvvm.cp_async_bulk_tensor_g2s_tile_2d_multicast` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d` from `int_nvvm_cp_async_bulk_tensor_g2s_tile_2d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [
], [
], [
], ;` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_multicast_cg2` | `nvvm.cp_async_bulk_tensor_g2s_tile_2d_multicast_cg2` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d` from `int_nvvm_cp_async_bulk_tensor_g2s_tile_2d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.6 on sm_100a or sm_101a or sm_103a or sm_110a ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.cta_group::2 [
], [
], [
], ;` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_2d_s2g` | `nvvm.cp_async_bulk_tensor_s2g_tile_2d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.2d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_2d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.2d.global.shared::cta.tile.bulk_group [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s` | `nvvm.cp_async_bulk_tensor_g2s_tile_3d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.tile.3d` from `int_nvvm_cp_async_bulk_tensor_g2s_tile_3d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | +| `cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s_cta` | `nvvm.cp_async_bulk_tensor_g2s_cta_tile_3d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.3d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_3d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.6 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma-cta`; expects `cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_3d_s2g` | `nvvm.cp_async_bulk_tensor_s2g_tile_3d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.3d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_3d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.3d.global.shared::cta.tile.bulk_group [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s` | `nvvm.cp_async_bulk_tensor_g2s_tile_4d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.tile.4d` from `int_nvvm_cp_async_bulk_tensor_g2s_tile_4d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | +| `cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s_cta` | `nvvm.cp_async_bulk_tensor_g2s_cta_tile_4d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.4d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_4d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.6 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma-cta`; expects `cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_4d_s2g` | `nvvm.cp_async_bulk_tensor_s2g_tile_4d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.4d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_4d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.4d.global.shared::cta.tile.bulk_group [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s` | `nvvm.cp_async_bulk_tensor_g2s_tile_5d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.tile.5d` from `int_nvvm_cp_async_bulk_tensor_g2s_tile_5d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | +| `cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s_cta` | `nvvm.cp_async_bulk_tensor_g2s_cta_tile_5d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.5d` from `int_nvvm_cp_async_bulk_tensor_g2s_cta_tile_5d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.6 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma-cta`; expects `cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes [
], [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_5d_s2g` | `nvvm.cp_async_bulk_tensor_s2g_tile_5d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.s2g.tile.5d` from `int_nvvm_cp_async_bulk_tensor_s2g_tile_5d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5 Asynchronous bulk tensor copy](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma`; expects `cp.async.bulk.tensor.5d.global.shared::cta.tile.bulk_group [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_reduce_add_im2col_3d` | `nvvm.cp_async_bulk_tensor_reduce_add_im2col_3d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.reduce.add.im2col.3d` from `int_nvvm_cp_async_bulk_tensor_reduce_add_im2col_3d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5.3 Data Movement and Conversion Instructions: cp.reduce.async.bulk.tensor](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-reduce-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma-reduce`; expects `cp.reduce.async.bulk.tensor.3d.global.shared::cta.add.im2col_no_offs.bulk_group [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | | `cuda_intrinsics::tma::cp_async_bulk_tensor_reduce_add_im2col_4d` | `nvvm.cp_async_bulk_tensor_reduce_add_im2col_4d` | LLVM `llvm.nvvm.cp.async.bulk.tensor.reduce.add.im2col.4d` from `int_nvvm_cp_async_bulk_tensor_reduce_add_im2col_4d` | unsafe; scope thread; impure; memory read_write; convergent true | PTX 8.0 on sm_90+ ([PTX ISA 9.3, 9.7.9.26.5.3 Data Movement and Conversion Instructions: cp.reduce.async.bulk.tensor](https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-cp-reduce-async-bulk-tensor)) | validated on `rust-llvm-23.1.3-aab11ee0-tma-reduce`; expects `cp.reduce.async.bulk.tensor.4d.global.shared::cta.add.im2col_no_offs.bulk_group [
], [
];` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) | @@ -1392,6 +1397,10 @@ This table is generated from the resolved catalog. Evidence stages below disting - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0328` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_1d_g2s` - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_1d_g2s` +- `cp_async_bulk_tensor_1d_g2s_cta` (`i1034`) + - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1034` + - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_1d_g2s_cta` + - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_1d_g2s_cta` - `cp_async_bulk_tensor_1d_s2g` (`i0335`) - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0335` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_1d_s2g` @@ -1400,6 +1409,10 @@ This table is generated from the resolved catalog. Evidence stages below disting - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0329` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s` - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_2d_g2s` +- `cp_async_bulk_tensor_2d_g2s_cta` (`i1035`) + - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1035` + - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_cta` + - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_2d_g2s_cta` - `cp_async_bulk_tensor_2d_g2s_multicast` (`i0330`) - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0330` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_2d_g2s_multicast` @@ -1416,6 +1429,10 @@ This table is generated from the resolved catalog. Evidence stages below disting - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0332` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s` - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_3d_g2s` +- `cp_async_bulk_tensor_3d_g2s_cta` (`i1036`) + - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1036` + - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_3d_g2s_cta` + - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_3d_g2s_cta` - `cp_async_bulk_tensor_3d_s2g` (`i0337`) - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0337` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_3d_s2g` @@ -1424,6 +1441,10 @@ This table is generated from the resolved catalog. Evidence stages below disting - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0333` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s` - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_4d_g2s` +- `cp_async_bulk_tensor_4d_g2s_cta` (`i1037`) + - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1037` + - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_4d_g2s_cta` + - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_4d_g2s_cta` - `cp_async_bulk_tensor_4d_s2g` (`i0338`) - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0338` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_4d_s2g` @@ -1432,6 +1453,10 @@ This table is generated from the resolved catalog. Evidence stages below disting - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0334` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s` - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_5d_g2s` +- `cp_async_bulk_tensor_5d_g2s_cta` (`i1038`) + - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i1038` + - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_5d_g2s_cta` + - compatibility compiler path: `cuda_device::tma::cp_async_bulk_tensor_5d_g2s_cta` - `cp_async_bulk_tensor_5d_s2g` (`i0339`) - canonical compiler path: `cuda_intrinsics::__cuda_oxide_intrinsic_abi_v1::i0339` - public source path: `cuda_intrinsics::tma::cp_async_bulk_tensor_5d_s2g` @@ -5827,16 +5852,21 @@ This table is generated from the resolved catalog. Evidence stages below disting - `cp_async_bulk_prefetch_tensor_gather4_2d_l2`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_TENSOR_PF_TILE_GATHER4_2D`; selection predicates `Subtarget->hasTMABlackwellSupport()` - `cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_TENSOR_PF_TILE_GATHER4_2D_CH`; selection predicates `Subtarget->hasTMABlackwellSupport()` - `cp_async_bulk_tensor_1d_g2s`: properties `ImmArg`, `ImmArg`, `ImmArg`, `IntrConvergent`, `Range`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_TILE_CG0_1D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` +- `cp_async_bulk_tensor_1d_g2s_cta`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_CTA_TILE_1D`; selection predicates `Subtarget->getPTXVersion() >= 86`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_1d_s2g`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_TENSOR_S2G_TILE_1D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_2d_g2s`: properties `ImmArg`, `ImmArg`, `ImmArg`, `IntrConvergent`, `Range`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_TILE_CG0_2D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` +- `cp_async_bulk_tensor_2d_g2s_cta`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_CTA_TILE_2D`; selection predicates `Subtarget->getPTXVersion() >= 86`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_2d_g2s_multicast`: properties `ImmArg`, `ImmArg`, `ImmArg`, `IntrConvergent`, `Range`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_TILE_CG0_2D_MC`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_2d_g2s_multicast_cg2`: properties `ImmArg`, `ImmArg`, `ImmArg`, `IntrConvergent`, `Range`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_TILE_2D_MC`; selection predicates `Subtarget->hasTMABlackwellSupport()` - `cp_async_bulk_tensor_2d_s2g`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_TENSOR_S2G_TILE_2D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_3d_g2s`: properties `ImmArg`, `ImmArg`, `ImmArg`, `IntrConvergent`, `Range`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_TILE_CG0_3D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` +- `cp_async_bulk_tensor_3d_g2s_cta`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_CTA_TILE_3D`; selection predicates `Subtarget->getPTXVersion() >= 86`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_3d_s2g`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_TENSOR_S2G_TILE_3D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_4d_g2s`: properties `ImmArg`, `ImmArg`, `ImmArg`, `IntrConvergent`, `Range`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_TILE_CG0_4D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` +- `cp_async_bulk_tensor_4d_g2s_cta`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_CTA_TILE_4D`; selection predicates `Subtarget->getPTXVersion() >= 86`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_4d_s2g`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_TENSOR_S2G_TILE_4D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_5d_g2s`: properties `ImmArg`, `ImmArg`, `ImmArg`, `IntrConvergent`, `Range`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_TILE_CG0_5D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` +- `cp_async_bulk_tensor_5d_g2s_cta`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `WriteOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_G2S_CTA_TILE_5D`; selection predicates `Subtarget->getPTXVersion() >= 86`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_5d_s2g`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records `TMA_TENSOR_S2G_TILE_5D`; selection predicates `Subtarget->getPTXVersion() >= 80`, `Subtarget->getSmVersion() >= 90` - `cp_async_bulk_tensor_reduce_add_im2col_3d`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records ; selection predicates none - `cp_async_bulk_tensor_reduce_add_im2col_4d`: properties `ImmArg`, `IntrConvergent`, `ReadOnly`, `ReadOnly`; result `NoUndef` false; half-open result range `none`; selection records ; selection predicates none @@ -7308,6 +7338,14 @@ This table is generated from the resolved catalog. Evidence stages below disting - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma` at PTX 8.0 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX. - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. +- `cp_async_bulk_tensor_1d_g2s_cta`: runtime `not recorded` + - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) + - declaration canonicalization on `sm_90`: `succeeded` — Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_1d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,007-byte source had SHA-256 253d3ea39f9d26f5000f83febb9f72a15af2bed6ad0f668ab206976ce4c6d6ba; the 2,640-byte bitcode had SHA-256 9f42944557579c5c9c4e38716cce52607a07c51e5f2272393948fdc40cfb8961; the 1,234-byte canonical IR had SHA-256 835c2efc3b53d5333ec6a8260fcad29d5d0ce889194d0523feb57f7b22281fff. + - backend codegen on `sm_90, ptx86`: `succeeded` — The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.1d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,165-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 b0b4469f60fea3342f58e26eb4698834c6cb0714aac17fd3ffaed79cbd4f226a. + - PTX assembly on `sm_90, ptx86`: `succeeded` — CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480. Tool `/usr/local/cuda-13.3/bin/ptxas` reports `Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0` (SHA-256 `afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc`). Artifact `cubin`. + - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) + - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60. + - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. - `cp_async_bulk_tensor_1d_s2g`: runtime `not recorded` - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma` at PTX 8.0 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) - declaration canonicalization on `sm_90`: `succeeded` — Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581. @@ -7324,6 +7362,14 @@ This table is generated from the resolved catalog. Evidence stages below disting - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma` at PTX 8.0 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX. - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. +- `cp_async_bulk_tensor_2d_g2s_cta`: runtime `not recorded` + - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) + - declaration canonicalization on `sm_90`: `succeeded` — Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_2d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,038-byte source had SHA-256 f0d47c1b6670854dfddce5bc45206d8f4ad16e9a05236dd4e1d4036b518666ea; the 2,648-byte bitcode had SHA-256 6ab53bfd2185eb8b6d7f7ffb3793d72cb7b48a07f61846a5317604700bdd3225; the 1,265-byte canonical IR had SHA-256 2c00613205683c093c5002f0f7347d3b2d6b82e46c5bc00072fb5a781f7d9d44. + - backend codegen on `sm_90, ptx86`: `succeeded` — The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.2d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,299-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 59fa9746f48c669cbff59faffea6039a796055f2721b8647a6331d78b5a26521. + - PTX assembly on `sm_90, ptx86`: `succeeded` — CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480. Tool `/usr/local/cuda-13.3/bin/ptxas` reports `Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0` (SHA-256 `afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc`). Artifact `cubin`. + - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) + - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60. + - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. - `cp_async_bulk_tensor_2d_g2s_multicast`: runtime `not recorded` - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma` at PTX 8.0 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) - declaration canonicalization on `sm_90`: `succeeded` — Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581. @@ -7356,6 +7402,14 @@ This table is generated from the resolved catalog. Evidence stages below disting - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma` at PTX 8.0 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX. - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. +- `cp_async_bulk_tensor_3d_g2s_cta`: runtime `not recorded` + - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) + - declaration canonicalization on `sm_90`: `succeeded` — Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_3d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,069-byte source had SHA-256 159ab50347ad4f81c8a110df7de35db35da9eb099938874523e5ec10ea867c8c; the 2,656-byte bitcode had SHA-256 3b5f19c93cffadf279efad33b501519f09cacbc660784f195ab3d007b5bca854; the 1,296-byte canonical IR had SHA-256 223dd701edae4ecf145ca18afc659dd5a50785cf8e492f6ba5c7b7f8299dc030. + - backend codegen on `sm_90, ptx86`: `succeeded` — The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.3d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,433-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 e9bab5f8e2c2a697ca435f86eaff468984a03470ae4bf28f975c2c89fc6676b9. + - PTX assembly on `sm_90, ptx86`: `succeeded` — CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480. Tool `/usr/local/cuda-13.3/bin/ptxas` reports `Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0` (SHA-256 `afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc`). Artifact `cubin`. + - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) + - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60. + - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. - `cp_async_bulk_tensor_3d_s2g`: runtime `not recorded` - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma` at PTX 8.0 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) - declaration canonicalization on `sm_90`: `succeeded` — Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581. @@ -7372,6 +7426,14 @@ This table is generated from the resolved catalog. Evidence stages below disting - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma` at PTX 8.0 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX. - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. +- `cp_async_bulk_tensor_4d_g2s_cta`: runtime `not recorded` + - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) + - declaration canonicalization on `sm_90`: `succeeded` — Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_4d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,100-byte source had SHA-256 becebd363b47d76e388fccf46646db4d8bba5225b4eb23cc7cae1673a5411ddd; the 2,664-byte bitcode had SHA-256 77f23ed44102d830fb8a576a8ff62f947e96b2b15871e2984083851d1e82f63e; the 1,327-byte canonical IR had SHA-256 6fd6a38abbb214d1e1fc5a4c8aa800ab963d1f53350afc3a7ec965501e467fb4. + - backend codegen on `sm_90, ptx86`: `succeeded` — The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.4d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,567-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 0607dfca39829f02dbda9367602cb27d28e6dc5dd8f5909a451fa4648442629a. + - PTX assembly on `sm_90, ptx86`: `succeeded` — CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480. Tool `/usr/local/cuda-13.3/bin/ptxas` reports `Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0` (SHA-256 `afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc`). Artifact `cubin`. + - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) + - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60. + - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. - `cp_async_bulk_tensor_4d_s2g`: runtime `not recorded` - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma` at PTX 8.0 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) - declaration canonicalization on `sm_90`: `succeeded` — Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581. @@ -7388,6 +7450,14 @@ This table is generated from the resolved catalog. Evidence stages below disting - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma` at PTX 8.0 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained every exact instruction with no extern calls. SHA-256 over the sorted per-operation ID/hash lists is 0b5d1f4be62738f2ec3f7c2d58a78bf4a984c225f07190229a422d2befcb71ce for source IR, fbb3c6200d5aab32ef48e87fe38ef428e0d44a8d1c99190ad4291a0a3a4f9797 for LTOIR, and 0f8b2b96f9dc634cb2e8c4ffdebff62f8bcebfe00ae08ba7ab2f975ad5ee382f for linked PTX. - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced valid CUDA ELF cubins; SHA-256 over the sorted per-operation ID/cubin-hash list is ea355e5985a293b10266691d71302aad755879854308203a8f9a3f3bd0ce1303. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. +- `cp_async_bulk_tensor_5d_g2s_cta`: runtime `not recorded` + - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) + - declaration canonicalization on `sm_90`: `succeeded` — Candidate probe (`probe --candidate --intrinsic cp_async_bulk_tensor_5d_g2s_cta --gpu-target sm_90 --ptx-feature +ptx86`) with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): llvm-as canonicalized the intrinsic attributes and the declaration facts matched the imported record. The 1,131-byte source had SHA-256 da8bf49e04c4250626e97150345c704f33802066582d2e80d1c17dd58c0fa0b0; the 2,668-byte bitcode had SHA-256 d1d535b89cc23b4983c8fb5b00562e06266454177d8eb3a81684859ac5430cb5; the 1,358-byte canonical IR had SHA-256 e4862fd8477461d79f65871ae6dc8750d273a0e50911602e0bfea91bae76df7c. + - backend codegen on `sm_90, ptx86`: `succeeded` — The pinned llc (-march=nvptx64 -mcpu=sm_90 -mattr=+ptx86) emitted the exact `cp.async.bulk.tensor.5d.shared::cta.global.tile.mbarrier::complete_tx::bytes` instruction with no extern declarations. The 1,701-byte PTX (`.version 8.6`, `.target sm_90`) had SHA-256 2f87ab7dc1ed68335e4ad29e4bcc7bd264779b33deb26f1d18e0c955f6bf8223. + - PTX assembly on `sm_90, ptx86`: `succeeded` — CUDA 13.3 ptxas (-arch=sm_90) produced a 1,720-byte cubin with SHA-256 9f49927d82f4aeb6e387fbbe19d851b010dd170a6a71ccf4b346d5cb49ecb480. Tool `/usr/local/cuda-13.3/bin/ptxas` reports `Cuda compilation tools, release 13.3, V13.3.73; build cuda_13.3.r13.3/compiler.38244171_0` (SHA-256 `afd8d1e1fa6e310f7faee44f6621e4c1315fb7fd6da7d4d87414358e12a651dc`). Artifact `cubin`. + - `lib_nvvm` uses `inline_ptx` from profile `cuda-13.3-libnvvm-13.3.73-tma-cta` at PTX 8.6 / sm_90+: status `validated` (`CUDA 13.3 libNVVM package 13.3.73; libNVVM API 2.0`, SHA-256 `5317a9fb3c8b719198e89e27fd60e2b59f2702fe6b1eab357167c920b52be9cc`) + - backend codegen on `compute_90, ptx93`: `succeeded` — CUDA 13.3.73 libNVVM retained all five exact 1D-5D instructions. Legacy NVVM IR SHA-256 fbde8ef7febd45940a6dbe906a5315dc8e90ad08d25d549bdfafa130e29071e6; LTOIR SHA-256 c08bf2d430ee33d995d17eaf7ab71fb419a9d23ed2db4423adb22230f2cce693; linked PTX SHA-256 26f53d690b7a281bc3cc5eedfb820858f7802f0428c8673d1012ff6fac18de60. + - device link on `sm_90, ptx93`: `succeeded` — CUDA 13.3 nvJitLink produced cubin SHA-256 87ec36d9e25d1c954361a691e300867c8b0c0e8f212aa228dd63adf640419cae; nvdisasm retained five TMA load instructions. Tool `libnvJitLink.so.13.3.33` reports `CUDA 13.3 nvJitLink 13.3.33` (SHA-256 `f84abd240b428825c3ef16632deb1026b208252fe19cef225a9f4ad06a39d1aa`). Artifact `cubin`. - `cp_async_bulk_tensor_5d_s2g`: runtime `not recorded` - `llvm_nvptx` uses `typed_nvvm` from profile `rust-llvm-23.1.3-aab11ee0-tma` at PTX 8.0 / sm_90+: status `validated` (`LLVM version 23.1.3-rust-1.101.0-nightly`, SHA-256 `7534b0236b9b26c7e6910548038b2285ee1bb5908e4700764dc78a7a1a54b449`) - declaration canonicalization on `sm_90`: `succeeded` — Revalidated 14 separate probes with nightly-2026-10-11 (rustc 6f5da406b56c2db6607202b63daee73aa8c61297; LLVM aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019): canonical typed declarations and their exact imported attributes passed. The sorted ID/artifact-kind/SHA-256 ledger in refresh-2026-10-11/stages.jsonl has SHA-256 efc480471cb427bf10e4f248a24454a98102aaef6bf200026684072b2f5f2581. diff --git a/cuda-oxide/intrinsics/overlay/tma.toml b/cuda-oxide/intrinsics/overlay/tma.toml index fc6867447..2e0115a0a 100644 --- a/cuda-oxide/intrinsics/overlay/tma.toml +++ b/cuda-oxide/intrinsics/overlay/tma.toml @@ -1,11 +1,13 @@ # Reviewed Tensor Memory Accelerator operations. -schema = 62 +schema = 65 family = "tma" [tma] llvm_evidence_profile = "rust-llvm-23.1.3-aab11ee0-tma" libnvvm_evidence_profile = "cuda-13.3-libnvvm-13.3.73-tma" +cta_llvm_evidence_profile = "rust-llvm-23.1.3-aab11ee0-tma-cta" +cta_libnvvm_evidence_profile = "cuda-13.3-libnvvm-13.3.73-tma-cta" reduce_llvm_evidence_profile = "rust-llvm-23.1.3-aab11ee0-tma-reduce" reduce_libnvvm_evidence_profile = "cuda-13.3-libnvvm-13.3.73-tma-reduce" runtime_validation = "unexecuted" @@ -14,10 +16,18 @@ runtime_validation = "unexecuted" abi_id = "i0328" operation = "g2s_tile1d" +[[tma.variant]] +abi_id = "i1034" +operation = "g2s_cta_tile1d" + [[tma.variant]] abi_id = "i0329" operation = "g2s_tile2d" +[[tma.variant]] +abi_id = "i1035" +operation = "g2s_cta_tile2d" + [[tma.variant]] abi_id = "i0330" operation = "g2s_tile2d_multicast" @@ -30,14 +40,26 @@ operation = "g2s_tile2d_multicast_cg2" abi_id = "i0332" operation = "g2s_tile3d" +[[tma.variant]] +abi_id = "i1036" +operation = "g2s_cta_tile3d" + [[tma.variant]] abi_id = "i0333" operation = "g2s_tile4d" +[[tma.variant]] +abi_id = "i1037" +operation = "g2s_cta_tile4d" + [[tma.variant]] abi_id = "i0334" operation = "g2s_tile5d" +[[tma.variant]] +abi_id = "i1038" +operation = "g2s_cta_tile5d" + [[tma.variant]] abi_id = "i0335" operation = "s2g_tile1d" diff --git a/cuda-oxide/intrinsics/probes/abs_bf16x2.ll b/cuda-oxide/intrinsics/probes/abs_bf16x2.ll index 067d4c64f..e46f6f406 100644 --- a/cuda-oxide/intrinsics/probes/abs_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/abs_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/abs_f16x2.ll b/cuda-oxide/intrinsics/probes/abs_f16x2.ll index 6dae85758..8d053af58 100644 --- a/cuda-oxide/intrinsics/probes/abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/active_mask.ll b/cuda-oxide/intrinsics/probes/active_mask.ll index 7fa13f210..53bb01071 100644 --- a/cuda-oxide/intrinsics/probes/active_mask.ll +++ b/cuda-oxide/intrinsics/probes/active_mask.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_bf16x2.ll b/cuda-oxide/intrinsics/probes/add_bf16x2.ll index 9a1878935..929185a05 100644 --- a/cuda-oxide/intrinsics/probes/add_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/add_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_f16x2.ll b/cuda-oxide/intrinsics/probes/add_f16x2.ll index fc6178157..1575fe048 100644 --- a/cuda-oxide/intrinsics/probes/add_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/add_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_f32x2.ll b/cuda-oxide/intrinsics/probes/add_f32x2.ll index 37d3007c2..8e9261908 100644 --- a/cuda-oxide/intrinsics/probes/add_f32x2.ll +++ b/cuda-oxide/intrinsics/probes/add_f32x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_ftz_f32x2.ll b/cuda-oxide/intrinsics/probes/add_ftz_f32x2.ll index 57b115569..d8fb76141 100644 --- a/cuda-oxide/intrinsics/probes/add_ftz_f32x2.ll +++ b/cuda-oxide/intrinsics/probes/add_ftz_f32x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rm_f32.ll b/cuda-oxide/intrinsics/probes/add_rm_f32.ll index 7052963bf..fd762bdaa 100644 --- a/cuda-oxide/intrinsics/probes/add_rm_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rm_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rm_f64.ll b/cuda-oxide/intrinsics/probes/add_rm_f64.ll index 93e78e764..d74c143bb 100644 --- a/cuda-oxide/intrinsics/probes/add_rm_f64.ll +++ b/cuda-oxide/intrinsics/probes/add_rm_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rm_ftz_f32.ll b/cuda-oxide/intrinsics/probes/add_rm_ftz_f32.ll index 0f4fb6b03..0ea771056 100644 --- a/cuda-oxide/intrinsics/probes/add_rm_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rm_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rm_ftz_sat_f32.ll b/cuda-oxide/intrinsics/probes/add_rm_ftz_sat_f32.ll index 90ee605e4..570c87425 100644 --- a/cuda-oxide/intrinsics/probes/add_rm_ftz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rm_ftz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rm_sat_f32.ll b/cuda-oxide/intrinsics/probes/add_rm_sat_f32.ll index d69fbd14a..931ce024d 100644 --- a/cuda-oxide/intrinsics/probes/add_rm_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rm_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rn_f32.ll b/cuda-oxide/intrinsics/probes/add_rn_f32.ll index 73ed7cd3b..638e681ce 100644 --- a/cuda-oxide/intrinsics/probes/add_rn_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rn_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rn_f64.ll b/cuda-oxide/intrinsics/probes/add_rn_f64.ll index 3a6786a3c..ae039497a 100644 --- a/cuda-oxide/intrinsics/probes/add_rn_f64.ll +++ b/cuda-oxide/intrinsics/probes/add_rn_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rn_ftz_f32.ll b/cuda-oxide/intrinsics/probes/add_rn_ftz_f32.ll index 9973ab1f5..465c651bb 100644 --- a/cuda-oxide/intrinsics/probes/add_rn_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rn_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rn_ftz_sat_f32.ll b/cuda-oxide/intrinsics/probes/add_rn_ftz_sat_f32.ll index 471162620..f3f552ca9 100644 --- a/cuda-oxide/intrinsics/probes/add_rn_ftz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rn_ftz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rn_sat_f32.ll b/cuda-oxide/intrinsics/probes/add_rn_sat_f32.ll index a8030adfb..ff3963dd1 100644 --- a/cuda-oxide/intrinsics/probes/add_rn_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rn_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rp_f32.ll b/cuda-oxide/intrinsics/probes/add_rp_f32.ll index 13a461bd4..23fd8ddad 100644 --- a/cuda-oxide/intrinsics/probes/add_rp_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rp_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rp_f64.ll b/cuda-oxide/intrinsics/probes/add_rp_f64.ll index dd49cf523..59d7d7003 100644 --- a/cuda-oxide/intrinsics/probes/add_rp_f64.ll +++ b/cuda-oxide/intrinsics/probes/add_rp_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rp_ftz_f32.ll b/cuda-oxide/intrinsics/probes/add_rp_ftz_f32.ll index 8cf544e9f..01a2de074 100644 --- a/cuda-oxide/intrinsics/probes/add_rp_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rp_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rp_ftz_sat_f32.ll b/cuda-oxide/intrinsics/probes/add_rp_ftz_sat_f32.ll index 6c24875b2..3c4177c83 100644 --- a/cuda-oxide/intrinsics/probes/add_rp_ftz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rp_ftz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rp_sat_f32.ll b/cuda-oxide/intrinsics/probes/add_rp_sat_f32.ll index 4fc32f3f8..a488949e2 100644 --- a/cuda-oxide/intrinsics/probes/add_rp_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rp_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rz_f32.ll b/cuda-oxide/intrinsics/probes/add_rz_f32.ll index 6ac13a58b..cacec2e59 100644 --- a/cuda-oxide/intrinsics/probes/add_rz_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rz_f64.ll b/cuda-oxide/intrinsics/probes/add_rz_f64.ll index 2af8a3023..cdcbdb81f 100644 --- a/cuda-oxide/intrinsics/probes/add_rz_f64.ll +++ b/cuda-oxide/intrinsics/probes/add_rz_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rz_ftz_f32.ll b/cuda-oxide/intrinsics/probes/add_rz_ftz_f32.ll index 2bcdc125c..ace477c68 100644 --- a/cuda-oxide/intrinsics/probes/add_rz_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rz_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rz_ftz_sat_f32.ll b/cuda-oxide/intrinsics/probes/add_rz_ftz_sat_f32.ll index 355e30175..0bce30539 100644 --- a/cuda-oxide/intrinsics/probes/add_rz_ftz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rz_ftz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/add_rz_sat_f32.ll b/cuda-oxide/intrinsics/probes/add_rz_sat_f32.ll index 094f9f6de..c5cbe3c4f 100644 --- a/cuda-oxide/intrinsics/probes/add_rz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/add_rz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/all_sync.ll b/cuda-oxide/intrinsics/probes/all_sync.ll index 29d10663d..7446ffeb9 100644 --- a/cuda-oxide/intrinsics/probes/all_sync.ll +++ b/cuda-oxide/intrinsics/probes/all_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/any_sync.ll b/cuda-oxide/intrinsics/probes/any_sync.ll index 576cf8dd3..19614abd1 100644 --- a/cuda-oxide/intrinsics/probes/any_sync.ll +++ b/cuda-oxide/intrinsics/probes/any_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ballot_sync.ll b/cuda-oxide/intrinsics/probes/ballot_sync.ll index 8befb7faa..c44d1b877 100644 --- a/cuda-oxide/intrinsics/probes/ballot_sync.ll +++ b/cuda-oxide/intrinsics/probes/ballot_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cluster_arrive.ll b/cuda-oxide/intrinsics/probes/barrier_cluster_arrive.ll index cd1c45339..e1eef9d7a 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cluster_arrive.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cluster_arrive.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_aligned.ll b/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_aligned.ll index 5199ada72..b4c112c16 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_aligned.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_aligned.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_relaxed.ll b/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_relaxed.ll index 2c4ba7b68..93b7a5a5b 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_relaxed.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_relaxed.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_relaxed_aligned.ll b/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_relaxed_aligned.ll index 2665f273a..5e5aa3947 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_relaxed_aligned.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cluster_arrive_relaxed_aligned.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cluster_wait.ll b/cuda-oxide/intrinsics/probes/barrier_cluster_wait.ll index e0304517a..8183f455a 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cluster_wait.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cluster_wait.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cluster_wait_aligned.ll b/cuda-oxide/intrinsics/probes/barrier_cluster_wait_aligned.ll index fb0f31906..e0adeef3e 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cluster_wait_aligned.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cluster_wait_aligned.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cta_arrive.ll b/cuda-oxide/intrinsics/probes/barrier_cta_arrive.ll index f527f256a..6ae0c40b8 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cta_arrive.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cta_arrive.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cta_arrive_aligned.ll b/cuda-oxide/intrinsics/probes/barrier_cta_arrive_aligned.ll index a79bf1924..413749a19 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cta_arrive_aligned.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cta_arrive_aligned.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cta_sync.ll b/cuda-oxide/intrinsics/probes/barrier_cta_sync.ll index b53080cc4..d59eb301f 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cta_sync.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cta_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/barrier_cta_sync_aligned.ll b/cuda-oxide/intrinsics/probes/barrier_cta_sync_aligned.ll index 8cfb36aa3..10c6dea36 100644 --- a/cuda-oxide/intrinsics/probes/barrier_cta_sync_aligned.ll +++ b/cuda-oxide/intrinsics/probes/barrier_cta_sync_aligned.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/block_dim_x.ll b/cuda-oxide/intrinsics/probes/block_dim_x.ll index e27224c43..8c1ca292c 100644 --- a/cuda-oxide/intrinsics/probes/block_dim_x.ll +++ b/cuda-oxide/intrinsics/probes/block_dim_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/block_dim_y.ll b/cuda-oxide/intrinsics/probes/block_dim_y.ll index 078e0407f..e766323f6 100644 --- a/cuda-oxide/intrinsics/probes/block_dim_y.ll +++ b/cuda-oxide/intrinsics/probes/block_dim_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/block_dim_z.ll b/cuda-oxide/intrinsics/probes/block_dim_z.ll index 328f912a0..51ac9520e 100644 --- a/cuda-oxide/intrinsics/probes/block_dim_z.ll +++ b/cuda-oxide/intrinsics/probes/block_dim_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/block_idx_x.ll b/cuda-oxide/intrinsics/probes/block_idx_x.ll index d3f10ee9b..ac4a5f858 100644 --- a/cuda-oxide/intrinsics/probes/block_idx_x.ll +++ b/cuda-oxide/intrinsics/probes/block_idx_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/block_idx_y.ll b/cuda-oxide/intrinsics/probes/block_idx_y.ll index 393af5214..16f2ef82f 100644 --- a/cuda-oxide/intrinsics/probes/block_idx_y.ll +++ b/cuda-oxide/intrinsics/probes/block_idx_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/block_idx_z.ll b/cuda-oxide/intrinsics/probes/block_idx_z.ll index 7e46684f4..bc5668cc3 100644 --- a/cuda-oxide/intrinsics/probes/block_idx_z.ll +++ b/cuda-oxide/intrinsics/probes/block_idx_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/breakpoint.ll b/cuda-oxide/intrinsics/probes/breakpoint.ll index 06a2c469b..4e802e281 100644 --- a/cuda-oxide/intrinsics/probes/breakpoint.ll +++ b/cuda-oxide/intrinsics/probes/breakpoint.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_x.ll b/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_x.ll index f743449a7..9d5559f08 100644 --- a/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_x.ll +++ b/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_y.ll b/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_y.ll index 03b30c409..521403f12 100644 --- a/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_y.ll +++ b/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_z.ll b/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_z.ll index f7e0cdfe7..3f1b74542 100644 --- a/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_z.ll +++ b/cuda-oxide/intrinsics/probes/clc_query_get_first_ctaid_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clc_query_is_canceled.ll b/cuda-oxide/intrinsics/probes/clc_query_is_canceled.ll index 3fb1cef4c..0ba56f6f2 100644 --- a/cuda-oxide/intrinsics/probes/clc_query_is_canceled.ll +++ b/cuda-oxide/intrinsics/probes/clc_query_is_canceled.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clc_try_cancel.ll b/cuda-oxide/intrinsics/probes/clc_try_cancel.ll index c4c08feb4..6a6e97132 100644 --- a/cuda-oxide/intrinsics/probes/clc_try_cancel.ll +++ b/cuda-oxide/intrinsics/probes/clc_try_cancel.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clc_try_cancel_multicast.ll b/cuda-oxide/intrinsics/probes/clc_try_cancel_multicast.ll index 68c4df50e..17b988450 100644 --- a/cuda-oxide/intrinsics/probes/clc_try_cancel_multicast.ll +++ b/cuda-oxide/intrinsics/probes/clc_try_cancel_multicast.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clock.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/clock.lib_nvvm.ll index 386bf7683..269ea615a 100644 --- a/cuda-oxide/intrinsics/probes/clock.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/clock.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clock.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/clock.llvm_nvptx.ll index 4c04d476d..58eee26aa 100644 --- a/cuda-oxide/intrinsics/probes/clock.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/clock.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clock64.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/clock64.lib_nvvm.ll index f29a9214a..814e8c1f1 100644 --- a/cuda-oxide/intrinsics/probes/clock64.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/clock64.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/clock64.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/clock64.llvm_nvptx.ll index 0c7b38daa..060131862 100644 --- a/cuda-oxide/intrinsics/probes/clock64.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/clock64.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_block_count.ll b/cuda-oxide/intrinsics/probes/cluster_block_count.ll index d244f044a..4096cec17 100644 --- a/cuda-oxide/intrinsics/probes/cluster_block_count.ll +++ b/cuda-oxide/intrinsics/probes/cluster_block_count.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_block_idx_x.ll b/cuda-oxide/intrinsics/probes/cluster_block_idx_x.ll index 7488d448a..d2103b891 100644 --- a/cuda-oxide/intrinsics/probes/cluster_block_idx_x.ll +++ b/cuda-oxide/intrinsics/probes/cluster_block_idx_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_block_idx_y.ll b/cuda-oxide/intrinsics/probes/cluster_block_idx_y.ll index 7376e4d98..1e63183e9 100644 --- a/cuda-oxide/intrinsics/probes/cluster_block_idx_y.ll +++ b/cuda-oxide/intrinsics/probes/cluster_block_idx_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_block_idx_z.ll b/cuda-oxide/intrinsics/probes/cluster_block_idx_z.ll index 1a2fc3478..987c3da4f 100644 --- a/cuda-oxide/intrinsics/probes/cluster_block_idx_z.ll +++ b/cuda-oxide/intrinsics/probes/cluster_block_idx_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_block_rank.ll b/cuda-oxide/intrinsics/probes/cluster_block_rank.ll index 440c88061..7613a6deb 100644 --- a/cuda-oxide/intrinsics/probes/cluster_block_rank.ll +++ b/cuda-oxide/intrinsics/probes/cluster_block_rank.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_dim_x.ll b/cuda-oxide/intrinsics/probes/cluster_dim_x.ll index 7230c4358..093965029 100644 --- a/cuda-oxide/intrinsics/probes/cluster_dim_x.ll +++ b/cuda-oxide/intrinsics/probes/cluster_dim_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_dim_y.ll b/cuda-oxide/intrinsics/probes/cluster_dim_y.ll index 7c4e5810c..8d4cfdd62 100644 --- a/cuda-oxide/intrinsics/probes/cluster_dim_y.ll +++ b/cuda-oxide/intrinsics/probes/cluster_dim_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_dim_z.ll b/cuda-oxide/intrinsics/probes/cluster_dim_z.ll index e7c273060..7ba508695 100644 --- a/cuda-oxide/intrinsics/probes/cluster_dim_z.ll +++ b/cuda-oxide/intrinsics/probes/cluster_dim_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_grid_dim_x.ll b/cuda-oxide/intrinsics/probes/cluster_grid_dim_x.ll index 1dd35af07..4a3244946 100644 --- a/cuda-oxide/intrinsics/probes/cluster_grid_dim_x.ll +++ b/cuda-oxide/intrinsics/probes/cluster_grid_dim_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_grid_dim_y.ll b/cuda-oxide/intrinsics/probes/cluster_grid_dim_y.ll index 9508406af..77e47ab58 100644 --- a/cuda-oxide/intrinsics/probes/cluster_grid_dim_y.ll +++ b/cuda-oxide/intrinsics/probes/cluster_grid_dim_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_grid_dim_z.ll b/cuda-oxide/intrinsics/probes/cluster_grid_dim_z.ll index 910ee9bc3..ac60470cf 100644 --- a/cuda-oxide/intrinsics/probes/cluster_grid_dim_z.ll +++ b/cuda-oxide/intrinsics/probes/cluster_grid_dim_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_idx_x.ll b/cuda-oxide/intrinsics/probes/cluster_idx_x.ll index 8e799974c..e1703a1e6 100644 --- a/cuda-oxide/intrinsics/probes/cluster_idx_x.ll +++ b/cuda-oxide/intrinsics/probes/cluster_idx_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_idx_y.ll b/cuda-oxide/intrinsics/probes/cluster_idx_y.ll index a445fe11c..a61c26449 100644 --- a/cuda-oxide/intrinsics/probes/cluster_idx_y.ll +++ b/cuda-oxide/intrinsics/probes/cluster_idx_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cluster_idx_z.ll b/cuda-oxide/intrinsics/probes/cluster_idx_z.ll index 64bab3b05..65a453dc5 100644 --- a/cuda-oxide/intrinsics/probes/cluster_idx_z.ll +++ b/cuda-oxide/intrinsics/probes/cluster_idx_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cos_approx_f32.ll b/cuda-oxide/intrinsics/probes/cos_approx_f32.ll index 17c45671b..18ed784b2 100644 --- a/cuda-oxide/intrinsics/probes/cos_approx_f32.ll +++ b/cuda-oxide/intrinsics/probes/cos_approx_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cos_approx_ftz_f32.ll b/cuda-oxide/intrinsics/probes/cos_approx_ftz_f32.ll index c00c249af..988d73e76 100644 --- a/cuda-oxide/intrinsics/probes/cos_approx_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/cos_approx_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_commit_group.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_commit_group.ll index aaec000da..e9978bd65 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_commit_group.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_commit_group.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_1d_l2.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_1d_l2.ll index 46e442df9..a547fb53f 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_1d_l2.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_1d_l2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_1d_l2_cache_hint.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_1d_l2_cache_hint.ll index 9f6592181..44ed1e361 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_1d_l2_cache_hint.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_1d_l2_cache_hint.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_2d_l2.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_2d_l2.ll index 454680f01..3a662ba40 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_2d_l2.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_2d_l2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_2d_l2_cache_hint.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_2d_l2_cache_hint.ll index 68cde3cd8..22fc3d808 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_2d_l2_cache_hint.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_2d_l2_cache_hint.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_3d_l2.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_3d_l2.ll index 3a129e2a1..89b6304b2 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_3d_l2.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_3d_l2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_3d_l2_cache_hint.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_3d_l2_cache_hint.ll index 2b171015e..202d5c91e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_3d_l2_cache_hint.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_3d_l2_cache_hint.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_4d_l2.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_4d_l2.ll index 739d1e76a..b142222fa 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_4d_l2.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_4d_l2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_4d_l2_cache_hint.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_4d_l2_cache_hint.ll index a7443c9ee..dcedc6093 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_4d_l2_cache_hint.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_4d_l2_cache_hint.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_5d_l2.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_5d_l2.ll index 6f1160d85..0b1031554 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_5d_l2.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_5d_l2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_5d_l2_cache_hint.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_5d_l2_cache_hint.ll index bf99c9742..bc32f8c21 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_5d_l2_cache_hint.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_5d_l2_cache_hint.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_gather4_2d_l2.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_gather4_2d_l2.ll index c1adbd156..140271b48 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_gather4_2d_l2.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_gather4_2d_l2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint.ll index f499d5ff0..d7f4619d5 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_prefetch_tensor_gather4_2d_l2_cache_hint.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_g2s.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_g2s.ll index 5f6cd7d1a..2448c8e8b 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_g2s.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_g2s.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_g2s_cta.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_g2s_cta.ll new file mode 100644 index 000000000..1c2365a73 --- /dev/null +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_g2s_cta.ll @@ -0,0 +1,18 @@ +; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. +; SPDX-License-Identifier: Apache-2.0 +; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. +; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. + +target triple = "nvptx64-nvidia-cuda" + +declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.1d(ptr addrspace(3), ptr addrspace(3), ptr, i32, i64, i1) #0 + +define void @probe_cp_async_bulk_tensor_1d_g2s_cta(ptr %dst_generic, ptr %barrier_generic, ptr %tensor_map, i32 %coord0) #0 { + %dst = addrspacecast ptr %dst_generic to ptr addrspace(3) + %barrier = addrspacecast ptr %barrier_generic to ptr addrspace(3) + call void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.1d(ptr addrspace(3) %dst, ptr addrspace(3) %barrier, ptr %tensor_map, i32 %coord0, i64 0, i1 false) #0 + ret void +} + +attributes #0 = { convergent } diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_s2g.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_s2g.ll index 26c100136..33459bfe6 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_s2g.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_1d_s2g.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s.ll index e5287efe7..4bba5ecfa 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_cta.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_cta.ll new file mode 100644 index 000000000..d1fa8e289 --- /dev/null +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_cta.ll @@ -0,0 +1,18 @@ +; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. +; SPDX-License-Identifier: Apache-2.0 +; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. +; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. + +target triple = "nvptx64-nvidia-cuda" + +declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d(ptr addrspace(3), ptr addrspace(3), ptr, i32, i32, i64, i1) #0 + +define void @probe_cp_async_bulk_tensor_2d_g2s_cta(ptr %dst_generic, ptr %barrier_generic, ptr %tensor_map, i32 %coord0, i32 %coord1) #0 { + %dst = addrspacecast ptr %dst_generic to ptr addrspace(3) + %barrier = addrspacecast ptr %barrier_generic to ptr addrspace(3) + call void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.2d(ptr addrspace(3) %dst, ptr addrspace(3) %barrier, ptr %tensor_map, i32 %coord0, i32 %coord1, i64 0, i1 false) #0 + ret void +} + +attributes #0 = { convergent } diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_multicast.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_multicast.ll index 8e432bfd6..f8410a175 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_multicast.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_multicast.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_multicast_cg2.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_multicast_cg2.ll index d4c22e1e0..299801aa2 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_multicast_cg2.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_g2s_multicast_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_s2g.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_s2g.ll index ff4a9d96c..7ee36fefb 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_s2g.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_2d_s2g.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_g2s.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_g2s.ll index f4e59a83b..5d569b93c 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_g2s.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_g2s.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_g2s_cta.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_g2s_cta.ll new file mode 100644 index 000000000..a8cedd535 --- /dev/null +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_g2s_cta.ll @@ -0,0 +1,18 @@ +; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. +; SPDX-License-Identifier: Apache-2.0 +; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. +; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. + +target triple = "nvptx64-nvidia-cuda" + +declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.3d(ptr addrspace(3), ptr addrspace(3), ptr, i32, i32, i32, i64, i1) #0 + +define void @probe_cp_async_bulk_tensor_3d_g2s_cta(ptr %dst_generic, ptr %barrier_generic, ptr %tensor_map, i32 %coord0, i32 %coord1, i32 %coord2) #0 { + %dst = addrspacecast ptr %dst_generic to ptr addrspace(3) + %barrier = addrspacecast ptr %barrier_generic to ptr addrspace(3) + call void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.3d(ptr addrspace(3) %dst, ptr addrspace(3) %barrier, ptr %tensor_map, i32 %coord0, i32 %coord1, i32 %coord2, i64 0, i1 false) #0 + ret void +} + +attributes #0 = { convergent } diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_s2g.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_s2g.ll index e360ec555..2df0bf6db 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_s2g.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_3d_s2g.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_g2s.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_g2s.ll index 734c32c34..57a76256e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_g2s.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_g2s.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_g2s_cta.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_g2s_cta.ll new file mode 100644 index 000000000..c80f62fe8 --- /dev/null +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_g2s_cta.ll @@ -0,0 +1,18 @@ +; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. +; SPDX-License-Identifier: Apache-2.0 +; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. +; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. + +target triple = "nvptx64-nvidia-cuda" + +declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.4d(ptr addrspace(3), ptr addrspace(3), ptr, i32, i32, i32, i32, i64, i1) #0 + +define void @probe_cp_async_bulk_tensor_4d_g2s_cta(ptr %dst_generic, ptr %barrier_generic, ptr %tensor_map, i32 %coord0, i32 %coord1, i32 %coord2, i32 %coord3) #0 { + %dst = addrspacecast ptr %dst_generic to ptr addrspace(3) + %barrier = addrspacecast ptr %barrier_generic to ptr addrspace(3) + call void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.4d(ptr addrspace(3) %dst, ptr addrspace(3) %barrier, ptr %tensor_map, i32 %coord0, i32 %coord1, i32 %coord2, i32 %coord3, i64 0, i1 false) #0 + ret void +} + +attributes #0 = { convergent } diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_s2g.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_s2g.ll index 78aa88746..bd156b2c6 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_s2g.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_4d_s2g.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_g2s.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_g2s.ll index 59c458104..7b8fbc58d 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_g2s.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_g2s.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_g2s_cta.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_g2s_cta.ll new file mode 100644 index 000000000..94386ac5b --- /dev/null +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_g2s_cta.ll @@ -0,0 +1,18 @@ +; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. +; SPDX-License-Identifier: Apache-2.0 +; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. +; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. + +target triple = "nvptx64-nvidia-cuda" + +declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.5d(ptr addrspace(3), ptr addrspace(3), ptr, i32, i32, i32, i32, i32, i64, i1) #0 + +define void @probe_cp_async_bulk_tensor_5d_g2s_cta(ptr %dst_generic, ptr %barrier_generic, ptr %tensor_map, i32 %coord0, i32 %coord1, i32 %coord2, i32 %coord3, i32 %coord4) #0 { + %dst = addrspacecast ptr %dst_generic to ptr addrspace(3) + %barrier = addrspacecast ptr %barrier_generic to ptr addrspace(3) + call void @llvm.nvvm.cp.async.bulk.tensor.g2s.cta.tile.5d(ptr addrspace(3) %dst, ptr addrspace(3) %barrier, ptr %tensor_map, i32 %coord0, i32 %coord1, i32 %coord2, i32 %coord3, i32 %coord4, i64 0, i1 false) #0 + ret void +} + +attributes #0 = { convergent } diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_s2g.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_s2g.ll index 08a7b47bf..f8d5c93c9 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_s2g.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_5d_s2g.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_3d.ll index a328c13b5..f6aa4172c 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_4d.ll index 47a90c1f9..b6bc08759 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_5d.ll index b440cc9a6..ca99453c1 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_im2col_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_1d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_1d.ll index 42c8c3406..bd956a1dd 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_1d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_1d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_2d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_2d.ll index af359d940..b81c4d98b 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_2d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_2d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_3d.ll index 4ad7a9bbd..cde371688 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_4d.ll index cdf181292..fd705ce50 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_5d.ll index 69068f434..6d6ee437e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_add_tile_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_3d.ll index 39b8079ff..8e2360145 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_4d.ll index a8488e950..fbb9d1b5b 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_5d.ll index e9db974bc..002ab772d 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_im2col_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_1d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_1d.ll index a1bfb7b66..f2127942d 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_1d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_1d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_2d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_2d.ll index e53fff68a..80d295962 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_2d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_2d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_3d.ll index 8edc1a796..3b1f925ad 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_4d.ll index 3454107c8..42a2b67ac 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_5d.ll index a93faef3d..97cd77243 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_and_tile_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_3d.ll index 81cd03fd9..6c9c30b6e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_4d.ll index 36750c05d..584b70188 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_5d.ll index 460a9bca6..ee9d4909b 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_im2col_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_1d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_1d.ll index 4e2cc0c1e..19c3cbae2 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_1d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_1d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_2d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_2d.ll index a4dcb9af5..f2fa71d72 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_2d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_2d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_3d.ll index 9911f2ff9..c894f8e03 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_4d.ll index 5d64d411f..01ed1f1b7 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_5d.ll index 636780c00..ab14acb4e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_dec_tile_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_3d.ll index fcfd9f8f7..cbbf3727e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_4d.ll index 851b12833..dedddbe4e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_5d.ll index 57eba85ca..4fcb85480 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_im2col_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_1d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_1d.ll index d44234722..86e44d8e2 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_1d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_1d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_2d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_2d.ll index 59875f1bd..2d9d2bb84 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_2d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_2d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_3d.ll index 9aea0c597..2ae37f44c 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_4d.ll index d1717d054..a6a36ac8f 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_5d.ll index 686b3a345..a37d3e946 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_inc_tile_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_3d.ll index 8ac11f8ad..eea7f519e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_4d.ll index 3eba84995..c91873dc0 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_5d.ll index 82af3065d..42b42a5df 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_im2col_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_1d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_1d.ll index 18f95c6b3..8cc06b0e0 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_1d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_1d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_2d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_2d.ll index e3f194e2e..f4f16b6d1 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_2d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_2d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_3d.ll index 5e9dc6726..0b4875b30 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_4d.ll index bab5881b9..4ac76bbda 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_5d.ll index 7ed9a51e1..82b9fb6e5 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_max_tile_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_3d.ll index b2f6a782c..b0d724dc1 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_4d.ll index e3ff3c3db..fe17ef8bf 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_5d.ll index 95f338523..e3eab2383 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_im2col_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_1d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_1d.ll index a0f458825..e449dabd0 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_1d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_1d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_2d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_2d.ll index d02c1e355..4a370d31e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_2d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_2d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_3d.ll index 83f57a71a..6a82737b5 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_4d.ll index adab16d05..1d1da9485 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_5d.ll index d5b4f9f0c..698b703b7 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_min_tile_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_3d.ll index feaf9c4b5..f6e79e25c 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_4d.ll index eb6f62f03..f7ccb958a 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_5d.ll index 2910ba2c1..1bc4a2c47 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_im2col_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_1d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_1d.ll index e07e14ea5..4f4bcc914 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_1d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_1d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_2d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_2d.ll index 87f260e83..76e9cfce0 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_2d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_2d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_3d.ll index 4fcf98c00..e84327ff9 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_4d.ll index e4ef06fc3..f0ebe0204 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_5d.ll index d803d51e0..3336bb6cb 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_or_tile_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_3d.ll index a885575b6..b6c04ef32 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_4d.ll index a8fdaadc1..7289e3d0e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_5d.ll index 4f20aabc3..49447c963 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_im2col_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_1d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_1d.ll index 72f78bf37..9f7da7d68 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_1d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_1d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_2d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_2d.ll index c57638e74..47112c5c5 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_2d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_2d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_3d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_3d.ll index 0b3816e9d..c9796b53b 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_3d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_3d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_4d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_4d.ll index 0313e9fcb..d44ec05c1 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_4d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_4d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_5d.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_5d.ll index 572c3abda..068480049 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_5d.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_tensor_reduce_xor_tile_5d.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_wait_group.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_wait_group.ll index fcb98f7e9..052f0e5f8 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_wait_group.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_wait_group.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_bulk_wait_group_read.ll b/cuda-oxide/intrinsics/probes/cp_async_bulk_wait_group_read.ll index ddd4e3021..30cf4e78e 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_bulk_wait_group_read.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_bulk_wait_group_read.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_ca_16.ll b/cuda-oxide/intrinsics/probes/cp_async_ca_16.ll index 92065d24e..c2a13457f 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_ca_16.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_ca_16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_ca_4.ll b/cuda-oxide/intrinsics/probes/cp_async_ca_4.ll index 152c11dce..8be6e32b1 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_ca_4.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_ca_4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_ca_8.ll b/cuda-oxide/intrinsics/probes/cp_async_ca_8.ll index 0e6b58d2b..cea3ee60f 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_ca_8.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_ca_8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_16.ll b/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_16.ll index 91eed5c73..6ced4b639 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_16.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_4.ll b/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_4.ll index 0b45e2547..164af51c2 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_4.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_8.ll b/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_8.ll index aa2bfb45a..5ed88a284 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_8.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_ca_zfill_8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_cg_16.ll b/cuda-oxide/intrinsics/probes/cp_async_cg_16.ll index 6b81d8bd5..da742891a 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_cg_16.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_cg_16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_cg_zfill_16.ll b/cuda-oxide/intrinsics/probes/cp_async_cg_zfill_16.ll index c6fb788be..798ebb5e6 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_cg_zfill_16.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_cg_zfill_16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_commit_group.ll b/cuda-oxide/intrinsics/probes/cp_async_commit_group.ll index 9ae3549b3..fe7d15e01 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_commit_group.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_commit_group.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive.ll b/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive.ll index 4255a9103..26df93f72 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_noinc.ll b/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_noinc.ll index dcbc9a3db..64838e7da 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_noinc.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_noinc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_noinc_shared.ll b/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_noinc_shared.ll index 41db65046..0af7393ba 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_noinc_shared.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_noinc_shared.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_shared.ll b/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_shared.ll index 66ce94af5..a666f1820 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_shared.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_mbarrier_arrive_shared.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_wait_all.ll b/cuda-oxide/intrinsics/probes/cp_async_wait_all.ll index 4429f3f90..afff37782 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_wait_all.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_wait_all.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cp_async_wait_group.ll b/cuda-oxide/intrinsics/probes/cp_async_wait_group.ll index ed8825d63..fad0504d4 100644 --- a/cuda-oxide/intrinsics/probes/cp_async_wait_group.ll +++ b/cuda-oxide/intrinsics/probes/cp_async_wait_group.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_f16x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_f16x2_f32.ll index e87cebd79..0493a146d 100644 --- a/cuda-oxide/intrinsics/probes/cvt_f16x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_f16x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_f32x2_bf16x2.ll b/cuda-oxide/intrinsics/probes/cvt_f32x2_bf16x2.ll index 64c4b60d4..b95ff64c5 100644 --- a/cuda-oxide/intrinsics/probes/cvt_f32x2_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_f32x2_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_f16x2_e4m3x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_f16x2_e4m3x2.ll index 7dbe4d58c..198b91f59 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_f16x2_e4m3x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_f16x2_e4m3x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_f16x2_e5m2x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_f16x2_e5m2x2.ll index 180687268..597d5e863 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_f16x2_e5m2x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_f16x2_e5m2x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_relu_bf16x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_relu_bf16x2_f32.ll index 9aaebadea..358d0d197 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_relu_bf16x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_relu_bf16x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_e4m3x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_e4m3x2.ll index 8a82e4f56..83a34555f 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_e4m3x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_e4m3x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_e5m2x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_e5m2x2.ll index 6109e6e7b..d1abfdbeb 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_e5m2x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_e5m2x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_f32.ll index ca58ea76c..fb4e380e0 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_relu_f16x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_relu_satfinite_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_relu_satfinite_tf32_f32.ll index e5e9348b8..19500ce50 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_relu_satfinite_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_relu_satfinite_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_relu_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_relu_tf32_f32.ll index e6a5e3436..a4a66fa8d 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_relu_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_relu_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_bf16x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_bf16x2.ll index 081544c23..0bb892f6d 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_f16x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_f16x2.ll index 5e934210d..df07680c3 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_f32.ll index 589183604..3f3eebf39 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e4m3x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_bf16x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_bf16x2.ll index 0bb0c2478..dff829ede 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_f16x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_f16x2.ll index f440693c2..ad47108f8 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_f32.ll index 1e7961cb6..d2b2ab9df 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_e5m2x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_bf16x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_bf16x2.ll index 0b467a2c9..8913f5b5c 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_f16x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_f16x2.ll index a286c6d1b..198f3643a 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_f32.ll index 7ab9c6003..aea882afb 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e4m3x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_bf16x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_bf16x2.ll index 299f50b53..1699d7f09 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_f16x2.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_f16x2.ll index 84aa86de9..87af035c3 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_f32.ll index cc500b731..5474c487d 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_relu_e5m2x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_tf32_f32.ll index 171314946..346a8f82b 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_satfinite_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rn_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rn_tf32_f32.ll index f47633313..d5949622e 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rn_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rn_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rna_satfinite_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rna_satfinite_tf32_f32.ll index 3229ba9b7..4998500d2 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rna_satfinite_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rna_satfinite_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rna_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rna_tf32_f32.ll index db51c2bab..72b07524e 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rna_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rna_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rz_bf16x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rz_bf16x2_f32.ll index b76bdde10..b07b8ae0c 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rz_bf16x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rz_bf16x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rz_f16x2_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rz_f16x2_f32.ll index fe7d3d693..d1b870520 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rz_f16x2_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rz_f16x2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rz_relu_satfinite_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rz_relu_satfinite_tf32_f32.ll index 82ffd726e..562d04368 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rz_relu_satfinite_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rz_relu_satfinite_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rz_relu_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rz_relu_tf32_f32.ll index 74f5fc2ba..1c9506aee 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rz_relu_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rz_relu_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rz_satfinite_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rz_satfinite_tf32_f32.ll index 63eef2db4..caf879709 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rz_satfinite_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rz_satfinite_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/cvt_rz_tf32_f32.ll b/cuda-oxide/intrinsics/probes/cvt_rz_tf32_f32.ll index 419b47bfc..dcde35438 100644 --- a/cuda-oxide/intrinsics/probes/cvt_rz_tf32_f32.ll +++ b/cuda-oxide/intrinsics/probes/cvt_rz_tf32_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rm_f32.ll b/cuda-oxide/intrinsics/probes/div_rm_f32.ll index a1d3da671..7a42bbe5d 100644 --- a/cuda-oxide/intrinsics/probes/div_rm_f32.ll +++ b/cuda-oxide/intrinsics/probes/div_rm_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rm_f64.ll b/cuda-oxide/intrinsics/probes/div_rm_f64.ll index 00c593b7c..29061f0f7 100644 --- a/cuda-oxide/intrinsics/probes/div_rm_f64.ll +++ b/cuda-oxide/intrinsics/probes/div_rm_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rm_ftz_f32.ll b/cuda-oxide/intrinsics/probes/div_rm_ftz_f32.ll index 0f51f288c..c7b2c9189 100644 --- a/cuda-oxide/intrinsics/probes/div_rm_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/div_rm_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rn_f32.ll b/cuda-oxide/intrinsics/probes/div_rn_f32.ll index da743c6ea..c80996935 100644 --- a/cuda-oxide/intrinsics/probes/div_rn_f32.ll +++ b/cuda-oxide/intrinsics/probes/div_rn_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rn_f64.ll b/cuda-oxide/intrinsics/probes/div_rn_f64.ll index 47f2055c6..fb1fb654a 100644 --- a/cuda-oxide/intrinsics/probes/div_rn_f64.ll +++ b/cuda-oxide/intrinsics/probes/div_rn_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rn_ftz_f32.ll b/cuda-oxide/intrinsics/probes/div_rn_ftz_f32.ll index 016ec9e93..29f7cb650 100644 --- a/cuda-oxide/intrinsics/probes/div_rn_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/div_rn_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rp_f32.ll b/cuda-oxide/intrinsics/probes/div_rp_f32.ll index 206e3ad10..7d0f0f9bf 100644 --- a/cuda-oxide/intrinsics/probes/div_rp_f32.ll +++ b/cuda-oxide/intrinsics/probes/div_rp_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rp_f64.ll b/cuda-oxide/intrinsics/probes/div_rp_f64.ll index 6d5d138bf..ed58974ab 100644 --- a/cuda-oxide/intrinsics/probes/div_rp_f64.ll +++ b/cuda-oxide/intrinsics/probes/div_rp_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rp_ftz_f32.ll b/cuda-oxide/intrinsics/probes/div_rp_ftz_f32.ll index 3433f00e9..22547b5bd 100644 --- a/cuda-oxide/intrinsics/probes/div_rp_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/div_rp_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rz_f32.ll b/cuda-oxide/intrinsics/probes/div_rz_f32.ll index 278d9fb16..630217b84 100644 --- a/cuda-oxide/intrinsics/probes/div_rz_f32.ll +++ b/cuda-oxide/intrinsics/probes/div_rz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rz_f64.ll b/cuda-oxide/intrinsics/probes/div_rz_f64.ll index 3369341ae..642e6bda8 100644 --- a/cuda-oxide/intrinsics/probes/div_rz_f64.ll +++ b/cuda-oxide/intrinsics/probes/div_rz_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/div_rz_ftz_f32.ll b/cuda-oxide/intrinsics/probes/div_rz_ftz_f32.ll index 447898907..2d2f5e51a 100644 --- a/cuda-oxide/intrinsics/probes/div_rz_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/div_rz_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/dp2a_s32.ll b/cuda-oxide/intrinsics/probes/dp2a_s32.ll index 33b9322ca..d9261cfed 100644 --- a/cuda-oxide/intrinsics/probes/dp2a_s32.ll +++ b/cuda-oxide/intrinsics/probes/dp2a_s32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/dp2a_u32.ll b/cuda-oxide/intrinsics/probes/dp2a_u32.ll index 5c4f90aac..ce6c83d58 100644 --- a/cuda-oxide/intrinsics/probes/dp2a_u32.ll +++ b/cuda-oxide/intrinsics/probes/dp2a_u32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/dp4a_s32.ll b/cuda-oxide/intrinsics/probes/dp4a_s32.ll index 8030171ce..2d41f630a 100644 --- a/cuda-oxide/intrinsics/probes/dp4a_s32.ll +++ b/cuda-oxide/intrinsics/probes/dp4a_s32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/dp4a_u32.ll b/cuda-oxide/intrinsics/probes/dp4a_u32.ll index 7259fd2a0..2572e6291 100644 --- a/cuda-oxide/intrinsics/probes/dp4a_u32.ll +++ b/cuda-oxide/intrinsics/probes/dp4a_u32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/dsmem_read_u32.ll b/cuda-oxide/intrinsics/probes/dsmem_read_u32.ll index fcd64174a..233350c6c 100644 --- a/cuda-oxide/intrinsics/probes/dsmem_read_u32.ll +++ b/cuda-oxide/intrinsics/probes/dsmem_read_u32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/dynamic_smem_size.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/dynamic_smem_size.lib_nvvm.ll index c0947ece4..965f3f0e2 100644 --- a/cuda-oxide/intrinsics/probes/dynamic_smem_size.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/dynamic_smem_size.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/dynamic_smem_size.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/dynamic_smem_size.llvm_nvptx.ll index b5b176f52..a92c49207 100644 --- a/cuda-oxide/intrinsics/probes/dynamic_smem_size.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/dynamic_smem_size.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/elect_sync.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/elect_sync.lib_nvvm.ll index b1113d808..d2057ada5 100644 --- a/cuda-oxide/intrinsics/probes/elect_sync.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/elect_sync.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/elect_sync.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/elect_sync.llvm_nvptx.ll index 27907c641..a3081473d 100644 --- a/cuda-oxide/intrinsics/probes/elect_sync.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/elect_sync.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/envreg1.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/envreg1.lib_nvvm.ll index 5a813a2d8..2c420d59b 100644 --- a/cuda-oxide/intrinsics/probes/envreg1.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/envreg1.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/envreg1.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/envreg1.llvm_nvptx.ll index 70ecb71ba..e405529d2 100644 --- a/cuda-oxide/intrinsics/probes/envreg1.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/envreg1.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/envreg2.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/envreg2.lib_nvvm.ll index 3f3050d23..2b118653f 100644 --- a/cuda-oxide/intrinsics/probes/envreg2.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/envreg2.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/envreg2.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/envreg2.llvm_nvptx.ll index e61b17ab3..3149169d4 100644 --- a/cuda-oxide/intrinsics/probes/envreg2.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/envreg2.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ex2_approx_f16.ll b/cuda-oxide/intrinsics/probes/ex2_approx_f16.ll index a19c13f31..3819cc7d1 100644 --- a/cuda-oxide/intrinsics/probes/ex2_approx_f16.ll +++ b/cuda-oxide/intrinsics/probes/ex2_approx_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ex2_approx_f32.ll b/cuda-oxide/intrinsics/probes/ex2_approx_f32.ll index 56f292188..d0191b77c 100644 --- a/cuda-oxide/intrinsics/probes/ex2_approx_f32.ll +++ b/cuda-oxide/intrinsics/probes/ex2_approx_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ex2_approx_ftz_f32.ll b/cuda-oxide/intrinsics/probes/ex2_approx_ftz_f32.ll index 64c080d15..5c3f9f3ef 100644 --- a/cuda-oxide/intrinsics/probes/ex2_approx_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/ex2_approx_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_mbarrier_init_release_cluster.ll b/cuda-oxide/intrinsics/probes/fence_mbarrier_init_release_cluster.ll index 21aa09cb0..ac6874666 100644 --- a/cuda-oxide/intrinsics/probes/fence_mbarrier_init_release_cluster.ll +++ b/cuda-oxide/intrinsics/probes/fence_mbarrier_init_release_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_async_generic_acquire_shared_cluster_cluster.ll b/cuda-oxide/intrinsics/probes/fence_proxy_async_generic_acquire_shared_cluster_cluster.ll index 0ac2068b5..307b76d16 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_async_generic_acquire_shared_cluster_cluster.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_async_generic_acquire_shared_cluster_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_async_generic_release_shared_cta_cluster.ll b/cuda-oxide/intrinsics/probes/fence_proxy_async_generic_release_shared_cta_cluster.ll index 5c2acf139..1261c0f26 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_async_generic_release_shared_cta_cluster.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_async_generic_release_shared_cta_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_async_shared_cta.ll b/cuda-oxide/intrinsics/probes/fence_proxy_async_shared_cta.ll index 49531a6d3..a2ade6ecd 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_async_shared_cta.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_async_shared_cta.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_cluster.ll b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_cluster.ll index e1422568b..ab1ee0448 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_cluster.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_cta.ll b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_cta.ll index 336e78599..f68edcdb4 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_cta.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_cta.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_gpu.ll b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_gpu.ll index 2403be316..49836c3e1 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_gpu.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_gpu.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_system.ll b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_system.ll index 49812021b..5400396c2 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_system.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_acquire_system.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_cluster.ll b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_cluster.ll index 85452598d..65531318e 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_cluster.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_cta.ll b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_cta.ll index d18301c77..d018df00b 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_cta.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_cta.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_gpu.ll b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_gpu.ll index 8c93d7571..6c2ea273e 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_gpu.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_gpu.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_system.ll b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_system.ll index 5e17dedb6..458bd1e5c 100644 --- a/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_system.ll +++ b/cuda-oxide/intrinsics/probes/fence_proxy_tensormap_generic_release_system.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_bf16x2.ll b/cuda-oxide/intrinsics/probes/fma_bf16x2.ll index 7ade7c929..f2dccf646 100644 --- a/cuda-oxide/intrinsics/probes/fma_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_f16x2.ll b/cuda-oxide/intrinsics/probes/fma_f16x2.ll index f98be54f1..a597f9614 100644 --- a/cuda-oxide/intrinsics/probes/fma_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_f32x2.ll b/cuda-oxide/intrinsics/probes/fma_f32x2.ll index dd332b10e..ae86d4e6d 100644 --- a/cuda-oxide/intrinsics/probes/fma_f32x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_f32x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_ftz_f16x2.ll b/cuda-oxide/intrinsics/probes/fma_ftz_f16x2.ll index 7053c4091..9ca84e44a 100644 --- a/cuda-oxide/intrinsics/probes/fma_ftz_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_ftz_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_ftz_f32x2.ll b/cuda-oxide/intrinsics/probes/fma_ftz_f32x2.ll index 67522a4e4..f2e3e561f 100644 --- a/cuda-oxide/intrinsics/probes/fma_ftz_f32x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_ftz_f32x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_ftz_relu_f16x2.ll b/cuda-oxide/intrinsics/probes/fma_ftz_relu_f16x2.ll index 8e3c9eb71..258330ba0 100644 --- a/cuda-oxide/intrinsics/probes/fma_ftz_relu_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_ftz_relu_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_ftz_sat_f16x2.ll b/cuda-oxide/intrinsics/probes/fma_ftz_sat_f16x2.ll index f0aa509ed..df131526f 100644 --- a/cuda-oxide/intrinsics/probes/fma_ftz_sat_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_ftz_sat_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_relu_bf16x2.ll b/cuda-oxide/intrinsics/probes/fma_relu_bf16x2.ll index ed3b27aec..748826136 100644 --- a/cuda-oxide/intrinsics/probes/fma_relu_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_relu_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_relu_f16x2.ll b/cuda-oxide/intrinsics/probes/fma_relu_f16x2.ll index 1c97d0f67..5584d04ed 100644 --- a/cuda-oxide/intrinsics/probes/fma_relu_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_relu_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rm_f32.ll b/cuda-oxide/intrinsics/probes/fma_rm_f32.ll index 92c934285..4af2cc618 100644 --- a/cuda-oxide/intrinsics/probes/fma_rm_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rm_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rm_f64.ll b/cuda-oxide/intrinsics/probes/fma_rm_f64.ll index 343fa5d58..a5f113a6f 100644 --- a/cuda-oxide/intrinsics/probes/fma_rm_f64.ll +++ b/cuda-oxide/intrinsics/probes/fma_rm_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rm_ftz_f32.ll b/cuda-oxide/intrinsics/probes/fma_rm_ftz_f32.ll index a6ad65b13..507e315a2 100644 --- a/cuda-oxide/intrinsics/probes/fma_rm_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rm_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rm_ftz_sat_f32.ll b/cuda-oxide/intrinsics/probes/fma_rm_ftz_sat_f32.ll index 6d25197ac..fb7eb628f 100644 --- a/cuda-oxide/intrinsics/probes/fma_rm_ftz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rm_ftz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rm_sat_f32.ll b/cuda-oxide/intrinsics/probes/fma_rm_sat_f32.ll index 075296a96..43aea59ea 100644 --- a/cuda-oxide/intrinsics/probes/fma_rm_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rm_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rn_f32.ll b/cuda-oxide/intrinsics/probes/fma_rn_f32.ll index 315fa265e..53840e35b 100644 --- a/cuda-oxide/intrinsics/probes/fma_rn_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rn_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rn_f64.ll b/cuda-oxide/intrinsics/probes/fma_rn_f64.ll index 9da6f952d..6712e072e 100644 --- a/cuda-oxide/intrinsics/probes/fma_rn_f64.ll +++ b/cuda-oxide/intrinsics/probes/fma_rn_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rn_ftz_f32.ll b/cuda-oxide/intrinsics/probes/fma_rn_ftz_f32.ll index aff9f07ff..cb48e916e 100644 --- a/cuda-oxide/intrinsics/probes/fma_rn_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rn_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rn_ftz_sat_f32.ll b/cuda-oxide/intrinsics/probes/fma_rn_ftz_sat_f32.ll index 5c471fec5..157102fd1 100644 --- a/cuda-oxide/intrinsics/probes/fma_rn_ftz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rn_ftz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rn_sat_f32.ll b/cuda-oxide/intrinsics/probes/fma_rn_sat_f32.ll index c8c4f68e4..7cee895c3 100644 --- a/cuda-oxide/intrinsics/probes/fma_rn_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rn_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rp_f32.ll b/cuda-oxide/intrinsics/probes/fma_rp_f32.ll index 399094b7f..b75d30555 100644 --- a/cuda-oxide/intrinsics/probes/fma_rp_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rp_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rp_f64.ll b/cuda-oxide/intrinsics/probes/fma_rp_f64.ll index 4213b63aa..15f6beff4 100644 --- a/cuda-oxide/intrinsics/probes/fma_rp_f64.ll +++ b/cuda-oxide/intrinsics/probes/fma_rp_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rp_ftz_f32.ll b/cuda-oxide/intrinsics/probes/fma_rp_ftz_f32.ll index 98cea80bd..925796f88 100644 --- a/cuda-oxide/intrinsics/probes/fma_rp_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rp_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rp_ftz_sat_f32.ll b/cuda-oxide/intrinsics/probes/fma_rp_ftz_sat_f32.ll index ae1ee6167..834e3a7f0 100644 --- a/cuda-oxide/intrinsics/probes/fma_rp_ftz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rp_ftz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rp_sat_f32.ll b/cuda-oxide/intrinsics/probes/fma_rp_sat_f32.ll index 7ca026128..4f2bacb79 100644 --- a/cuda-oxide/intrinsics/probes/fma_rp_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rp_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rz_f32.ll b/cuda-oxide/intrinsics/probes/fma_rz_f32.ll index 365dea58a..ab50a01c5 100644 --- a/cuda-oxide/intrinsics/probes/fma_rz_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rz_f64.ll b/cuda-oxide/intrinsics/probes/fma_rz_f64.ll index 0e5ec433d..559ac118f 100644 --- a/cuda-oxide/intrinsics/probes/fma_rz_f64.ll +++ b/cuda-oxide/intrinsics/probes/fma_rz_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rz_ftz_f32.ll b/cuda-oxide/intrinsics/probes/fma_rz_ftz_f32.ll index 3188b4513..2ee08e700 100644 --- a/cuda-oxide/intrinsics/probes/fma_rz_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rz_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rz_ftz_sat_f32.ll b/cuda-oxide/intrinsics/probes/fma_rz_ftz_sat_f32.ll index 3a0c7b627..ff741ba69 100644 --- a/cuda-oxide/intrinsics/probes/fma_rz_ftz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rz_ftz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_rz_sat_f32.ll b/cuda-oxide/intrinsics/probes/fma_rz_sat_f32.ll index 9ca05876f..46e5127db 100644 --- a/cuda-oxide/intrinsics/probes/fma_rz_sat_f32.ll +++ b/cuda-oxide/intrinsics/probes/fma_rz_sat_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/fma_sat_f16x2.ll b/cuda-oxide/intrinsics/probes/fma_sat_f16x2.ll index 78b9c47da..a4df1fd84 100644 --- a/cuda-oxide/intrinsics/probes/fma_sat_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/fma_sat_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/globaltimer.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/globaltimer.lib_nvvm.ll index 7d351266f..562de5263 100644 --- a/cuda-oxide/intrinsics/probes/globaltimer.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/globaltimer.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/globaltimer.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/globaltimer.llvm_nvptx.ll index c204923f7..cd4828591 100644 --- a/cuda-oxide/intrinsics/probes/globaltimer.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/globaltimer.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/grid_dependency_launch_dependents.ll b/cuda-oxide/intrinsics/probes/grid_dependency_launch_dependents.ll index 5b0764b9c..55f6b5b50 100644 --- a/cuda-oxide/intrinsics/probes/grid_dependency_launch_dependents.ll +++ b/cuda-oxide/intrinsics/probes/grid_dependency_launch_dependents.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/grid_dependency_wait.ll b/cuda-oxide/intrinsics/probes/grid_dependency_wait.ll index 6c1202bfa..2cbcfa0ee 100644 --- a/cuda-oxide/intrinsics/probes/grid_dependency_wait.ll +++ b/cuda-oxide/intrinsics/probes/grid_dependency_wait.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/grid_dim_x.ll b/cuda-oxide/intrinsics/probes/grid_dim_x.ll index 07aa2ea0b..4821d0c54 100644 --- a/cuda-oxide/intrinsics/probes/grid_dim_x.ll +++ b/cuda-oxide/intrinsics/probes/grid_dim_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/grid_dim_y.ll b/cuda-oxide/intrinsics/probes/grid_dim_y.ll index ad96d8b8d..47e3eaded 100644 --- a/cuda-oxide/intrinsics/probes/grid_dim_y.ll +++ b/cuda-oxide/intrinsics/probes/grid_dim_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/grid_dim_z.ll b/cuda-oxide/intrinsics/probes/grid_dim_z.ll index c28d2b2cc..572a3a5e0 100644 --- a/cuda-oxide/intrinsics/probes/grid_dim_z.ll +++ b/cuda-oxide/intrinsics/probes/grid_dim_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/gridid.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/gridid.lib_nvvm.ll index 4d46855c6..998302974 100644 --- a/cuda-oxide/intrinsics/probes/gridid.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/gridid.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/gridid.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/gridid.llvm_nvptx.ll index 4fdc43eb1..4b16a33b7 100644 --- a/cuda-oxide/intrinsics/probes/gridid.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/gridid.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/lane_id.ll b/cuda-oxide/intrinsics/probes/lane_id.ll index 50f6c834c..bb9bd2802 100644 --- a/cuda-oxide/intrinsics/probes/lane_id.ll +++ b/cuda-oxide/intrinsics/probes/lane_id.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/lanemask_eq.ll b/cuda-oxide/intrinsics/probes/lanemask_eq.ll index 61843c33c..f9b99d629 100644 --- a/cuda-oxide/intrinsics/probes/lanemask_eq.ll +++ b/cuda-oxide/intrinsics/probes/lanemask_eq.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/lanemask_ge.ll b/cuda-oxide/intrinsics/probes/lanemask_ge.ll index 1acedaa62..2a3fdf9be 100644 --- a/cuda-oxide/intrinsics/probes/lanemask_ge.ll +++ b/cuda-oxide/intrinsics/probes/lanemask_ge.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/lanemask_gt.ll b/cuda-oxide/intrinsics/probes/lanemask_gt.ll index 1295da46e..94c7501f6 100644 --- a/cuda-oxide/intrinsics/probes/lanemask_gt.ll +++ b/cuda-oxide/intrinsics/probes/lanemask_gt.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/lanemask_le.ll b/cuda-oxide/intrinsics/probes/lanemask_le.ll index e1c82790f..3dd6c7873 100644 --- a/cuda-oxide/intrinsics/probes/lanemask_le.ll +++ b/cuda-oxide/intrinsics/probes/lanemask_le.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/lanemask_lt.ll b/cuda-oxide/intrinsics/probes/lanemask_lt.ll index bb04d2313..01d1c2b18 100644 --- a/cuda-oxide/intrinsics/probes/lanemask_lt.ll +++ b/cuda-oxide/intrinsics/probes/lanemask_lt.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8.ll index d6c0fc68a..4594d1b13 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8x16_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8x16_b4x16_p64.ll index ddc3b57af..843d96fe5 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8x16_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8x16_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8x16_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8x16_b6x16_p32.ll index 3c9723fc8..330c2ee70 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8x16_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x1_trans_b8x16_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8.ll index 9c2ee7fca..56fceee0b 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8x16_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8x16_b4x16_p64.ll index e2fa0648b..984abf265 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8x16_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8x16_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8x16_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8x16_b6x16_p32.ll index c27457afd..4a9c597af 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8x16_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m16n16_x2_trans_b8x16_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x1_b8x16_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x1_b8x16_b4x16_p64.ll index 56bda5388..639ee3c0d 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x1_b8x16_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x1_b8x16_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x1_b8x16_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x1_b8x16_b6x16_p32.ll index 027688ac4..34972cac9 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x1_b8x16_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x1_b8x16_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x2_b8x16_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x2_b8x16_b4x16_p64.ll index 12ffeeed2..f386d7cc4 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x2_b8x16_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x2_b8x16_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x2_b8x16_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x2_b8x16_b6x16_p32.ll index 809f7982c..fbf88e2d4 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x2_b8x16_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x2_b8x16_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x4_b8x16_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x4_b8x16_b4x16_p64.ll index 6de8a3e41..5a016ff86 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x4_b8x16_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x4_b8x16_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x4_b8x16_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x4_b8x16_b6x16_p32.ll index 894afe32f..bccbf6014 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x4_b8x16_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n16_x4_b8x16_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x1_b16.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x1_b16.ll index 453adbec6..99c994fa3 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x1_b16.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x1_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x1_trans_b16.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x1_trans_b16.ll index 53cf01df1..f16610cd1 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x1_trans_b16.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x1_trans_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x2_b16.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x2_b16.ll index f96143f6e..686baf70a 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x2_b16.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x2_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x2_trans_b16.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x2_trans_b16.ll index 8d7a76b4a..f0ca18e74 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x2_trans_b16.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x2_trans_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x4_b16.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x4_b16.ll index dc7eb370f..4a56256f4 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x4_b16.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x4_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x4_trans_b16.ll b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x4_trans_b16.ll index d677b6c37..648ea122c 100644 --- a/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x4_trans_b16.ll +++ b/cuda-oxide/intrinsics/probes/ldmatrix_m8n8_x4_trans_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/lg2_approx_f32.ll b/cuda-oxide/intrinsics/probes/lg2_approx_f32.ll index e47328e66..a1653bc39 100644 --- a/cuda-oxide/intrinsics/probes/lg2_approx_f32.ll +++ b/cuda-oxide/intrinsics/probes/lg2_approx_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/lg2_approx_ftz_f32.ll b/cuda-oxide/intrinsics/probes/lg2_approx_ftz_f32.ll index fb08af5e0..cb0cbc19f 100644 --- a/cuda-oxide/intrinsics/probes/lg2_approx_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/lg2_approx_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/map_shared_rank.ll b/cuda-oxide/intrinsics/probes/map_shared_rank.ll index 5358a9d50..9b0480188 100644 --- a/cuda-oxide/intrinsics/probes/map_shared_rank.ll +++ b/cuda-oxide/intrinsics/probes/map_shared_rank.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/match_all_i64_sync.ll b/cuda-oxide/intrinsics/probes/match_all_i64_sync.ll index fdd5192dc..b18d94d21 100644 --- a/cuda-oxide/intrinsics/probes/match_all_i64_sync.ll +++ b/cuda-oxide/intrinsics/probes/match_all_i64_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/match_all_sync.ll b/cuda-oxide/intrinsics/probes/match_all_sync.ll index 868dd4db4..74b22b2ed 100644 --- a/cuda-oxide/intrinsics/probes/match_all_sync.ll +++ b/cuda-oxide/intrinsics/probes/match_all_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/match_any_i64_sync.ll b/cuda-oxide/intrinsics/probes/match_any_i64_sync.ll index bd6701cd5..0cadc0b65 100644 --- a/cuda-oxide/intrinsics/probes/match_any_i64_sync.ll +++ b/cuda-oxide/intrinsics/probes/match_any_i64_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/match_any_sync.ll b/cuda-oxide/intrinsics/probes/match_any_sync.ll index d130a1e5c..9109f2ce0 100644 --- a/cuda-oxide/intrinsics/probes/match_any_sync.ll +++ b/cuda-oxide/intrinsics/probes/match_any_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_bf16.ll b/cuda-oxide/intrinsics/probes/max_bf16.ll index 4c4301746..c19d49b1e 100644 --- a/cuda-oxide/intrinsics/probes/max_bf16.ll +++ b/cuda-oxide/intrinsics/probes/max_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_bf16x2.ll b/cuda-oxide/intrinsics/probes/max_bf16x2.ll index 4d54c4f92..2e4bc4e8a 100644 --- a/cuda-oxide/intrinsics/probes/max_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_f16.ll b/cuda-oxide/intrinsics/probes/max_f16.ll index 0bc3bc500..3644e944c 100644 --- a/cuda-oxide/intrinsics/probes/max_f16.ll +++ b/cuda-oxide/intrinsics/probes/max_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_f16x2.ll b/cuda-oxide/intrinsics/probes/max_f16x2.ll index 1ee7701c9..70deb4493 100644 --- a/cuda-oxide/intrinsics/probes/max_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_f16.ll b/cuda-oxide/intrinsics/probes/max_ftz_f16.ll index db80c0b0d..9d6d1c279 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_f16.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_f16x2.ll b/cuda-oxide/intrinsics/probes/max_ftz_f16x2.ll index 910871704..9adcdb614 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_nan_f16.ll b/cuda-oxide/intrinsics/probes/max_ftz_nan_f16.ll index 1a2b90b3b..04cbeab5a 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_nan_f16.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_nan_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_nan_f16x2.ll b/cuda-oxide/intrinsics/probes/max_ftz_nan_f16x2.ll index 009bc5b38..18d0996c4 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_nan_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_nan_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f16.ll b/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f16.ll index b0d5d6da4..8a5db4b16 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f16.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f16x2.ll b/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f16x2.ll index b082a62ef..ff20cce14 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f32.ll b/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f32.ll index 2725f06bb..effd7b96f 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_nan_xorsign_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f16.ll b/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f16.ll index 97a59e57e..08757fde1 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f16.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f16x2.ll b/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f16x2.ll index 6f366cd93..75480a758 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f32.ll b/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f32.ll index 8775a6961..0a9d03bab 100644 --- a/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/max_ftz_xorsign_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_bf16.ll b/cuda-oxide/intrinsics/probes/max_nan_bf16.ll index 915471f87..3c729d598 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_bf16.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_bf16x2.ll b/cuda-oxide/intrinsics/probes/max_nan_bf16x2.ll index 3aa7dab90..ba1284fa4 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_f16.ll b/cuda-oxide/intrinsics/probes/max_nan_f16.ll index 8bf3acc56..2c28b8539 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_f16.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_f16x2.ll b/cuda-oxide/intrinsics/probes/max_nan_f16x2.ll index 408b2a979..16ae09478 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_bf16.ll b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_bf16.ll index f32ff88d1..13b106899 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_bf16.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_bf16x2.ll b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_bf16x2.ll index f1c1aba39..4a27cdd08 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f16.ll b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f16.ll index 74a5ac743..0c293df2b 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f16.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f16x2.ll b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f16x2.ll index 776b6820e..787c60572 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f32.ll b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f32.ll index 7d64f9b05..a83d7b162 100644 --- a/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/max_nan_xorsign_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_relu_s16x2.ll b/cuda-oxide/intrinsics/probes/max_relu_s16x2.ll index 26a3d223c..e64f1fe37 100644 --- a/cuda-oxide/intrinsics/probes/max_relu_s16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_relu_s16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_relu_s32.ll b/cuda-oxide/intrinsics/probes/max_relu_s32.ll index d94f6d242..479a66e29 100644 --- a/cuda-oxide/intrinsics/probes/max_relu_s32.ll +++ b/cuda-oxide/intrinsics/probes/max_relu_s32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_s16x2.ll b/cuda-oxide/intrinsics/probes/max_s16x2.ll index 5a7722e61..d23da75a8 100644 --- a/cuda-oxide/intrinsics/probes/max_s16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_s16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_u16x2.ll b/cuda-oxide/intrinsics/probes/max_u16x2.ll index c5d4473cd..6c635e634 100644 --- a/cuda-oxide/intrinsics/probes/max_u16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_u16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_xorsign_abs_bf16.ll b/cuda-oxide/intrinsics/probes/max_xorsign_abs_bf16.ll index f1ce7e9a8..45ed4e48e 100644 --- a/cuda-oxide/intrinsics/probes/max_xorsign_abs_bf16.ll +++ b/cuda-oxide/intrinsics/probes/max_xorsign_abs_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_xorsign_abs_bf16x2.ll b/cuda-oxide/intrinsics/probes/max_xorsign_abs_bf16x2.ll index e2a400474..dbbda7340 100644 --- a/cuda-oxide/intrinsics/probes/max_xorsign_abs_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_xorsign_abs_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_xorsign_abs_f16.ll b/cuda-oxide/intrinsics/probes/max_xorsign_abs_f16.ll index bf91eb204..1c99646f4 100644 --- a/cuda-oxide/intrinsics/probes/max_xorsign_abs_f16.ll +++ b/cuda-oxide/intrinsics/probes/max_xorsign_abs_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_xorsign_abs_f16x2.ll b/cuda-oxide/intrinsics/probes/max_xorsign_abs_f16x2.ll index 8f9aae237..2c8610d38 100644 --- a/cuda-oxide/intrinsics/probes/max_xorsign_abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/max_xorsign_abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/max_xorsign_abs_f32.ll b/cuda-oxide/intrinsics/probes/max_xorsign_abs_f32.ll index 6389bee8a..e2b2fee03 100644 --- a/cuda-oxide/intrinsics/probes/max_xorsign_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/max_xorsign_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_arrive.ll b/cuda-oxide/intrinsics/probes/mbarrier_arrive.ll index a56003abd..7977c6a44 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_arrive.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_arrive.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_arrive_cluster.ll b/cuda-oxide/intrinsics/probes/mbarrier_arrive_cluster.ll index 51de19fe1..6aa112925 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_arrive_cluster.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_arrive_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_arrive_expect_tx.ll b/cuda-oxide/intrinsics/probes/mbarrier_arrive_expect_tx.ll index 736bb73f9..2cc015f97 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_arrive_expect_tx.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_arrive_expect_tx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_arrive_expect_tx_cluster.ll b/cuda-oxide/intrinsics/probes/mbarrier_arrive_expect_tx_cluster.ll index cf1247a80..419f1f94f 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_arrive_expect_tx_cluster.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_arrive_expect_tx_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_arrive_no_complete.ll b/cuda-oxide/intrinsics/probes/mbarrier_arrive_no_complete.ll index 6210d6381..8f1a6f7d7 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_arrive_no_complete.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_arrive_no_complete.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_init.ll b/cuda-oxide/intrinsics/probes/mbarrier_init.ll index 047b17588..ddea1c84e 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_init.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_init.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_inval.ll b/cuda-oxide/intrinsics/probes/mbarrier_inval.ll index baae5bf77..960026a0a 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_inval.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_inval.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_test_wait.ll b/cuda-oxide/intrinsics/probes/mbarrier_test_wait.ll index 0bbf79996..8090d9d56 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_test_wait.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_test_wait.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_try_wait.ll b/cuda-oxide/intrinsics/probes/mbarrier_try_wait.ll index 31af9e74b..3e52b13de 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_try_wait.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_try_wait.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_try_wait_parity.ll b/cuda-oxide/intrinsics/probes/mbarrier_try_wait_parity.ll index b4896de30..9c06fede4 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_try_wait_parity.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_try_wait_parity.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mbarrier_try_wait_parity_cluster.ll b/cuda-oxide/intrinsics/probes/mbarrier_try_wait_parity_cluster.ll index 6c9bdaeab..c3b4a9fe7 100644 --- a/cuda-oxide/intrinsics/probes/mbarrier_try_wait_parity_cluster.ll +++ b/cuda-oxide/intrinsics/probes/mbarrier_try_wait_parity_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_bf16.ll b/cuda-oxide/intrinsics/probes/min_bf16.ll index 14ed965fc..b3fd7c4ac 100644 --- a/cuda-oxide/intrinsics/probes/min_bf16.ll +++ b/cuda-oxide/intrinsics/probes/min_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_bf16x2.ll b/cuda-oxide/intrinsics/probes/min_bf16x2.ll index 2c6070ac9..620b7b956 100644 --- a/cuda-oxide/intrinsics/probes/min_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_f16.ll b/cuda-oxide/intrinsics/probes/min_f16.ll index 404f6960c..2ad6d6bf9 100644 --- a/cuda-oxide/intrinsics/probes/min_f16.ll +++ b/cuda-oxide/intrinsics/probes/min_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_f16x2.ll b/cuda-oxide/intrinsics/probes/min_f16x2.ll index f876d127d..e6e359ff2 100644 --- a/cuda-oxide/intrinsics/probes/min_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_f16.ll b/cuda-oxide/intrinsics/probes/min_ftz_f16.ll index bfc0cd2f6..1febefaa3 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_f16.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_f16x2.ll b/cuda-oxide/intrinsics/probes/min_ftz_f16x2.ll index 85d02ee23..19b8e9114 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_nan_f16.ll b/cuda-oxide/intrinsics/probes/min_ftz_nan_f16.ll index 50132f327..0b18d5af1 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_nan_f16.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_nan_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_nan_f16x2.ll b/cuda-oxide/intrinsics/probes/min_ftz_nan_f16x2.ll index 3481f3f28..32d672fce 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_nan_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_nan_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f16.ll b/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f16.ll index 06b403ec3..84a9df17e 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f16.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f16x2.ll b/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f16x2.ll index 7b1ebde1f..11652e1a8 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f32.ll b/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f32.ll index dbd6dd62f..f3e3f3231 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_nan_xorsign_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f16.ll b/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f16.ll index 958a82600..25075ee27 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f16.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f16x2.ll b/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f16x2.ll index 7c2e8409f..37ba2174d 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f32.ll b/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f32.ll index 2e1827fad..39b58cfed 100644 --- a/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/min_ftz_xorsign_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_bf16.ll b/cuda-oxide/intrinsics/probes/min_nan_bf16.ll index c110bc3aa..c2fea4168 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_bf16.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_bf16x2.ll b/cuda-oxide/intrinsics/probes/min_nan_bf16x2.ll index c3a58dec0..40e598ecc 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_f16.ll b/cuda-oxide/intrinsics/probes/min_nan_f16.ll index f8192f093..3069deaeb 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_f16.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_f16x2.ll b/cuda-oxide/intrinsics/probes/min_nan_f16x2.ll index f6fac96c8..f47c8d050 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_bf16.ll b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_bf16.ll index 75ff0178d..e1e7a35ad 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_bf16.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_bf16x2.ll b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_bf16x2.ll index b9b99edca..ff72d766b 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f16.ll b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f16.ll index 57cf031d1..124cba8c7 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f16.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f16x2.ll b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f16x2.ll index 8d7e74621..a0b989fc6 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f32.ll b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f32.ll index a798962b2..d7af7d6af 100644 --- a/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/min_nan_xorsign_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_relu_s16x2.ll b/cuda-oxide/intrinsics/probes/min_relu_s16x2.ll index c27c25059..de329394a 100644 --- a/cuda-oxide/intrinsics/probes/min_relu_s16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_relu_s16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_relu_s32.ll b/cuda-oxide/intrinsics/probes/min_relu_s32.ll index cc7b6317d..7b41a21a7 100644 --- a/cuda-oxide/intrinsics/probes/min_relu_s32.ll +++ b/cuda-oxide/intrinsics/probes/min_relu_s32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_s16x2.ll b/cuda-oxide/intrinsics/probes/min_s16x2.ll index cd633cc6d..fcf15664e 100644 --- a/cuda-oxide/intrinsics/probes/min_s16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_s16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_u16x2.ll b/cuda-oxide/intrinsics/probes/min_u16x2.ll index f41b0c332..8401b6d74 100644 --- a/cuda-oxide/intrinsics/probes/min_u16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_u16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_xorsign_abs_bf16.ll b/cuda-oxide/intrinsics/probes/min_xorsign_abs_bf16.ll index f7bde0891..af95cb911 100644 --- a/cuda-oxide/intrinsics/probes/min_xorsign_abs_bf16.ll +++ b/cuda-oxide/intrinsics/probes/min_xorsign_abs_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_xorsign_abs_bf16x2.ll b/cuda-oxide/intrinsics/probes/min_xorsign_abs_bf16x2.ll index 416660597..71f72dd73 100644 --- a/cuda-oxide/intrinsics/probes/min_xorsign_abs_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_xorsign_abs_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_xorsign_abs_f16.ll b/cuda-oxide/intrinsics/probes/min_xorsign_abs_f16.ll index 64bf809f5..44959208e 100644 --- a/cuda-oxide/intrinsics/probes/min_xorsign_abs_f16.ll +++ b/cuda-oxide/intrinsics/probes/min_xorsign_abs_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_xorsign_abs_f16x2.ll b/cuda-oxide/intrinsics/probes/min_xorsign_abs_f16x2.ll index b2c308038..b45e2130f 100644 --- a/cuda-oxide/intrinsics/probes/min_xorsign_abs_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/min_xorsign_abs_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/min_xorsign_abs_f32.ll b/cuda-oxide/intrinsics/probes/min_xorsign_abs_f32.ll index f8510d069..cc3f671f6 100644 --- a/cuda-oxide/intrinsics/probes/min_xorsign_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/min_xorsign_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k128_s32_b1_and_popc.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k128_s32_b1_and_popc.ll index 5fddf2f97..fa1372578 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k128_s32_b1_and_popc.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k128_s32_b1_and_popc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k128_s32_b1_xor_popc.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k128_s32_b1_xor_popc.ll index ef355d479..721525112 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k128_s32_b1_xor_popc.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k128_s32_b1_xor_popc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_f16_f16.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_f16_f16.ll index 0926220c4..1f191d26f 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_f16_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_f16_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_f32_bf16.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_f32_bf16.ll index bf74d189b..0ec720950 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_f32_bf16.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_f32_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_f32_f16.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_f32_f16.ll index 77fda8021..57e293c53 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_f32_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_f32_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e4m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e4m3_e4m3.ll index dfcd67fa9..325cec17e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e4m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e4m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e4m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e4m3_e5m2.ll index 7cfbe1497..ad7db3476 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e4m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e4m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e5m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e5m2_e4m3.ll index 194677aa5..b052cfc68 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e5m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e5m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e5m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e5m2_e5m2.ll index fb49b36c7..0d17ea05b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e5m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f16_e5m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e4m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e4m3_e4m3.ll index 6535323bb..f0934d659 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e4m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e4m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e4m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e4m3_e5m2.ll index f32e24c49..083649872 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e4m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e4m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e5m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e5m2_e4m3.ll index cebb70f36..44447744b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e5m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e5m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e5m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e5m2_e5m2.ll index b5d157fdc..a4db5c8b3 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e5m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_fp8_f32_e5m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8.ll index 60d0e7860..99cb12537 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_satfinite.ll index 6546046d1..d955d29da 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_u8.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_u8.ll index 383bd2770..18b20d47f 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_u8_satfinite.ll index 4d1a732b1..556ce9334 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_s8_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8.ll index e5dc531dd..9e6026e72 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_s8.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_s8.ll index ba122b4e6..f63c7a54d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_s8_satfinite.ll index 03bee05bd..079f93f28 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_satfinite.ll index dc1b9bbf0..846de7004 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k16_s32_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k256_s32_b1_and_popc.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k256_s32_b1_and_popc.ll index b5201afde..f15247556 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k256_s32_b1_and_popc.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k256_s32_b1_and_popc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k256_s32_b1_xor_popc.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k256_s32_b1_xor_popc.ll index 476a4038d..896a73419 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k256_s32_b1_xor_popc.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k256_s32_b1_xor_popc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e2m1.ll index 16f22fd9d..278f9db38 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e2m3.ll index 694133d6f..fae8a3a2b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e3m2.ll index 1a3a7b705..ce07d8c25 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e4m3.ll index df6658f4b..547d0666e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e5m2.ll index abcedf9ca..cce6fbeb8 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m1_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e2m1.ll index 7e993480e..027f63567 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e2m3.ll index ba8a3d239..d32c8a5ad 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e3m2.ll index d2f6566bc..118181a3a 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e4m3.ll index cb5873149..3bdcd6248 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e5m2.ll index e46dc1550..3a760376d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e2m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e2m1.ll index d15a414e4..617fe96f0 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e2m3.ll index 9e9620c22..3240530dc 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e3m2.ll index c5d2cf811..4abdd4682 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e4m3.ll index 65eaf27db..69ee1937a 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e5m2.ll index 93e80f881..691133df8 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e3m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e2m1.ll index 0c38aec93..afc4c3e7e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e2m3.ll index b1b3a4b4d..b4838e6e0 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e3m2.ll index 9d04dc8c9..a3880915b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e4m3.ll index 56313495d..c52534c95 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e5m2.ll index 204d1b2f7..21c9ec4ed 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e4m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e2m1.ll index e48d800b4..062cf1335 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e2m3.ll index cc3b2056c..d94339408 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e3m2.ll index e9a53d864..8bafc78b6 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e4m3.ll index bbba2ff0b..69751c224 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e5m2.ll index ae1653eb5..0ff2b9137 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f16_e5m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e2m1.ll index dba4a512b..a021186e6 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e2m3.ll index a15fa7dbe..3025a683a 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e3m2.ll index 0d75ea593..25b03cded 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e4m3.ll index 092f54ee2..c0cefaba8 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e5m2.ll index a8af62e48..a5e6f953d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m1_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e2m1.ll index d00fa9896..37d1f536e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e2m3.ll index e3a158096..2b64a954e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e3m2.ll index cbac0e377..e063b4b4d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e4m3.ll index 1ca4d0914..246ce8813 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e5m2.ll index e86249b58..f3f51e6e8 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e2m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e2m1.ll index bc87ee24c..a8c429622 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e2m3.ll index 091a7c67b..b935ff1e6 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e3m2.ll index 95b534220..b56b2123d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e4m3.ll index 4785e99b8..7a67c7b13 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e5m2.ll index 42547cf2d..759a09823 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e3m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e2m1.ll index 3d165cbae..180157808 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e2m3.ll index 811dcce5d..3721de21b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e3m2.ll index ff6d6cd68..61482b944 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e4m3.ll index 27d37e69c..4e18ef9af 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e5m2.ll index 63b97ffcd..99869663c 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e4m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e2m1.ll index 508a2510c..c45dbeaab 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e2m3.ll index 8b913fd68..e304d8bf2 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e3m2.ll index 3e337b5e1..498fec6b0 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e4m3.ll index 56e515743..773c74e1c 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e5m2.ll index 352fd5d01..22e1d45cf 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_f32_e5m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e4m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e4m3_e4m3.ll index 0e07abce1..68b769eea 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e4m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e4m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e4m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e4m3_e5m2.ll index c444bd307..51fac6397 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e4m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e4m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e5m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e5m2_e4m3.ll index 985769e69..6d33a1b21 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e5m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e5m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e5m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e5m2_e5m2.ll index 21487f325..f0b61fa73 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e5m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f16_e5m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e4m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e4m3_e4m3.ll index 8ef69b5e0..d1adfc3fc 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e4m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e4m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e4m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e4m3_e5m2.ll index 99351c9bf..b6e48f464 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e4m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e4m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e5m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e5m2_e4m3.ll index c71d6021e..da1b2375c 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e5m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e5m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e5m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e5m2_e5m2.ll index 95abd6f97..4ebd37f67 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e5m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_fp8_f32_e5m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e2m1.ll index 71a61bd42..1313d8220 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e2m3.ll index 984b30922..9a41b862b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e3m2.ll index fc6f40f9c..9c3351e8d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e4m3.ll index b89004ed2..8b571f3b4 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e5m2.ll index d8cccb3c8..db991e012 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m1_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e2m1.ll index c5b129b07..912c4dace 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e2m3.ll index 9f26f91a4..c96c256e6 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e3m2.ll index 9c8ed180b..734a4fc3e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e4m3.ll index e12f17e2c..7b40096d1 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e5m2.ll index 7b90b57cd..7a67ef3d0 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e2m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e2m1.ll index 42350b544..1bd1e836f 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e2m3.ll index 4fd25d377..31210f64c 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e3m2.ll index 2bd447cdc..14b473d0f 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e4m3.ll index 0a7db968b..da7b23119 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e5m2.ll index 36a4777c6..055104fb9 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e3m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e2m1.ll index 662d835a9..e85884472 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e2m3.ll index bab546777..327d254b1 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e3m2.ll index fd7ff404d..1bff57a99 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e4m3.ll index 2d7af4dc6..9da54c118 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e5m2.ll index eff1df7d5..a87b8c09b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e4m3_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e2m1.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e2m1.ll index f8b0ca442..9eb36193c 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e2m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e2m3.ll index 075addc2f..49c707ef9 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e3m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e3m2.ll index 74cb16cfb..b16213116 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e4m3.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e4m3.ll index d4288b006..f52667064 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e5m2.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e5m2.ll index d7a93b513..9bb17c4ff 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_mxf8f6f4_f32_e5m2_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4.ll index 2b05e6ac4..33d0d1ae4 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_satfinite.ll index a6d9df46f..398c97425 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_u4.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_u4.ll index 88978c792..6623b8076 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_u4_satfinite.ll index 7347863db..d6442d8e7 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s4_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8.ll index 1b82a45cb..36ba3bd35 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_satfinite.ll index 7f456cee1..d84f39e5d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_u8.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_u8.ll index c48bc4588..19c70460e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_u8_satfinite.ll index 86d578415..09b1e760f 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_s8_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4.ll index 100d8d02f..3b9a7607a 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_s4.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_s4.ll index 8fb5e99e1..ef5056fbb 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_s4_satfinite.ll index 363c34aa5..c1d361057 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_satfinite.ll index 0a4700d71..17ac02172 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8.ll index 1cc31f091..cdf865ea3 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_s8.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_s8.ll index a8f3e8edc..9f08c150b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_s8_satfinite.ll index b75a96d9c..9ea53896e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_satfinite.ll index e8cddafd2..139f6f4e5 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k32_s32_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k4_f32_tf32.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k4_f32_tf32.ll index df0a653fb..954476490 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k4_f32_tf32.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k4_f32_tf32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4.ll index 991065316..125f74e0f 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_satfinite.ll index cc6ad92c9..9ebe5599f 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_u4.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_u4.ll index bd201010c..0edb124ad 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_u4_satfinite.ll index fc5351eaa..c9893ef9a 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_s4_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4.ll index 2cd84c241..d09c98451 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_s4.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_s4.ll index e9d3682d8..193e5eb2f 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_s4_satfinite.ll index 45e334cca..08299ee17 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_satfinite.ll index 94995327a..51f174569 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k64_s32_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k8_f16_f16.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k8_f16_f16.ll index 4ddf6656e..607ea2d3e 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k8_f16_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k8_f16_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_bf16.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_bf16.ll index ca1249742..4470b8063 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_bf16.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_f16.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_f16.ll index 31875b3d5..7b083623d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_tf32.ll b/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_tf32.ll index ff8cf0fd4..8249863e9 100644 --- a/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_tf32.ll +++ b/cuda-oxide/intrinsics/probes/mma_m16n8k8_f32_tf32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k128_s32_b1_and_popc.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k128_s32_b1_and_popc.ll index c2c2b9235..aacee5dd9 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k128_s32_b1_and_popc.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k128_s32_b1_and_popc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k128_s32_b1_xor_popc.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k128_s32_b1_xor_popc.ll index 1703de2ea..3be88b3fb 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k128_s32_b1_xor_popc.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k128_s32_b1_xor_popc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8.ll index 56df23b76..729c2519b 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_satfinite.ll index 3f2d45dfa..e104fd966 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_u8.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_u8.ll index 88b613308..17f0a5559 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_u8_satfinite.ll index 35397f102..5266e35ee 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_s8_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8.ll index f64e01c53..93f5345f5 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_s8.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_s8.ll index 393c6a46e..c04b818f8 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_s8_satfinite.ll index a62d22771..8b536f8c0 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_satfinite.ll index 4e084fdc6..fa46c03dd 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k16_s32_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4.ll index 5779d1fa1..9ee6e89f9 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_satfinite.ll index 9ef79c8b3..667ebdfe6 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_u4.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_u4.ll index c5d0fa268..7e3b35fd0 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_u4_satfinite.ll index 8f121eec8..8f9ff8d84 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_s4_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4.ll index e5a8b93a3..07888edb9 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_s4.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_s4.ll index 75221e54a..aa2bed3bb 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_s4_satfinite.ll index 021fb9509..d7b5fbcbe 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_satfinite.ll index f0b419a65..82832000d 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k32_s32_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_m8n8k4_f64.ll b/cuda-oxide/intrinsics/probes/mma_m8n8k4_f64.ll index 3e3e7712c..1299aa0a7 100644 --- a/cuda-oxide/intrinsics/probes/mma_m8n8k4_f64.ll +++ b/cuda-oxide/intrinsics/probes/mma_m8n8k4_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4.ll index 8077529cd..7440bc82e 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_satfinite.ll index aaf3101ba..0a5fec376 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_u4.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_u4.ll index cbf2e9dcb..ae1dafa6a 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_u4_satfinite.ll index 6201f118e..329468af5 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_s4_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4.ll index 62a20ced1..55039ec6b 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_s4.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_s4.ll index 24f0d33c7..2cb509120 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_s4_satfinite.ll index 5f39597b9..cfe30cad8 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_satfinite.ll index df71dc6d0..6d58ec174 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k128_s32_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8.ll index 481a6cdfe..f24fabd66 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_satfinite.ll index 2dc97ff95..e952ce44c 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_u8.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_u8.ll index 37e2562d4..a60f5cb65 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_u8_satfinite.ll index 476882e1b..4d64a8341 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_s8_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8.ll index 8e7f54710..6d0e34be7 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_s8.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_s8.ll index 95c75878b..e4a784e0d 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_s8_satfinite.ll index eeb659840..3c23f9485 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_satfinite.ll index 4b32e8fc8..5c645518a 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k32_s32_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e4m3_e4m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e4m3_e4m3_f32.ll index 7207939f4..b24157208 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e4m3_e4m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e4m3_e4m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e4m3_e5m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e4m3_e5m2_f32.ll index ad54063aa..01a91d874 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e4m3_e5m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e4m3_e5m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e5m2_e4m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e5m2_e4m3_f32.ll index c126f9151..0fa2e8450 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e5m2_e4m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e5m2_e4m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e5m2_e5m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e5m2_e5m2_f32.ll index c9c1918d9..2f8d3c4f5 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e5m2_e5m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_f32_e5m2_e5m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4.ll index 619e06bc1..be2cc7c91 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_satfinite.ll index 37e8a6f05..3c04a35cc 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_u4.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_u4.ll index c85c9292a..5804d21c0 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_u4_satfinite.ll index 943cfced4..12c8dd5c5 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s4_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8.ll index 64891baec..e34ca13f9 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_satfinite.ll index ef1f8cb45..17bd289fd 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_u8.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_u8.ll index 935033518..836850432 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_u8_satfinite.ll index 2af2dbbea..d3755de6f 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_s8_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4.ll index 0be448972..50ffdace1 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_s4.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_s4.ll index 50df175b7..28429bb56 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_s4_satfinite.ll index 1f5e2846e..8cfa70d25 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_satfinite.ll index c82d70105..f1446edd8 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8.ll index e64308251..6f18f3517 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_s8.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_s8.ll index 4938e384b..cd6fc278c 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_s8_satfinite.ll index 1f7d24771..051c6d4c7 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_satfinite.ll index 11616cf36..4e1efacab 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_m16n8k64_s32_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4.ll index 7f2fab77e..800335d2a 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_satfinite.ll index 61c63143b..c390c28b1 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_u4.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_u4.ll index 8379d7c7c..9f3db5d2b 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_u4_satfinite.ll index a0f4f44d7..6cb31a9de 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_s4_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4.ll index 536ceec7d..786329924 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_s4.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_s4.ll index 0fd960de5..6c1e92fd4 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_s4_satfinite.ll index f4d0c1e71..6368aed6a 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_satfinite.ll index ac08ff4cb..1d35a1e47 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k128_s32_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f16_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f16_f16.ll index d1889712d..a731331f2 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f16_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f16_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_bf16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_bf16.ll index 82b0cbfca..2f3082d7b 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_bf16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_f16.ll index 6d183e3ed..211dbd4eb 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_tf32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_tf32.ll index 64459b199..7788ead15 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_tf32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k16_f32_tf32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f16_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f16_f16.ll index cd041ca53..4b4d733d8 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f16_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f16_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f32_bf16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f32_bf16.ll index eaa8c5c98..77f306baa 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f32_bf16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f32_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f32_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f32_f16.ll index e38d992dc..30182c35b 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f32_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_f32_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8.ll index c5274b850..ee5a01691 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_satfinite.ll index 4b76de039..f4d49a28b 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_u8.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_u8.ll index c68144b19..fe0ba54f6 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_u8_satfinite.ll index 6b07fa97b..6efea8eed 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_s8_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8.ll index 53b8ebb68..46cb773cd 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_s8.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_s8.ll index 4625dd516..a7b2e75a4 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_s8_satfinite.ll index fbd0f9893..bbe7abaf5 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_satfinite.ll index d0f994010..787d86058 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k32_s32_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e2m1_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e2m1_f16.ll index 18c39df4d..98a81aac1 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e2m1_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e2m1_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e2m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e2m3_f16.ll index badd9009b..a2a3a6753 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e2m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e2m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e3m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e3m2_f16.ll index 295478746..6124bc0e8 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e3m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e3m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e4m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e4m3_f16.ll index e4f5f34ec..01d7dc623 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e4m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e4m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e5m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e5m2_f16.ll index d7cdfc332..11c5b6454 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e5m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m1_e5m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e2m1_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e2m1_f16.ll index 49ab833b4..238b2faf0 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e2m1_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e2m1_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e2m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e2m3_f16.ll index 5f5e552bf..dac4744a8 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e2m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e2m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e3m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e3m2_f16.ll index 61216b1c2..7361c5ab4 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e3m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e3m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e4m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e4m3_f16.ll index 37e2d843d..95d4284f8 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e4m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e4m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e5m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e5m2_f16.ll index e355988fa..224116045 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e5m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e2m3_e5m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e2m1_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e2m1_f16.ll index a48ae8460..af0bd28c7 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e2m1_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e2m1_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e2m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e2m3_f16.ll index 365d99861..4e6bd73ae 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e2m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e2m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e3m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e3m2_f16.ll index 0c10aa3c5..2e0db1ea6 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e3m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e3m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e4m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e4m3_f16.ll index 7d2bf0ecd..7397268a7 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e4m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e4m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e5m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e5m2_f16.ll index 6d20bc300..c9f7763c5 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e5m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e3m2_e5m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e2m1_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e2m1_f16.ll index 04bcd8194..e8983ac1d 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e2m1_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e2m1_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e2m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e2m3_f16.ll index dbc72511a..440d5f6a0 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e2m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e2m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e3m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e3m2_f16.ll index 9bf94d05b..98939aca6 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e3m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e3m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e4m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e4m3_f16.ll index 744ede663..119503d4d 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e4m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e4m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e5m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e5m2_f16.ll index d169bcc77..6216ee701 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e5m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e4m3_e5m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e2m1_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e2m1_f16.ll index 896c6513a..e7a9b99f2 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e2m1_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e2m1_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e2m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e2m3_f16.ll index 9921ae107..509d491dd 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e2m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e2m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e3m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e3m2_f16.ll index 3ad855515..e1eaf9e21 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e3m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e3m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e4m3_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e4m3_f16.ll index 584298de5..84819143e 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e4m3_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e4m3_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e5m2_f16.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e5m2_f16.ll index ed9e47ebd..7eee7b9de 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e5m2_f16.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f16_e5m2_e5m2_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e2m1_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e2m1_f32.ll index 4b94ed7fa..f4f492388 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e2m1_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e2m1_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e2m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e2m3_f32.ll index cb687c945..250f13256 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e2m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e2m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e3m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e3m2_f32.ll index fa7b4aae5..c7c9ca16c 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e3m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e3m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e4m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e4m3_f32.ll index f925ccf29..e8d06477d 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e4m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e4m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e5m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e5m2_f32.ll index 338a03564..988f1d980 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e5m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m1_e5m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e2m1_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e2m1_f32.ll index 523615ae7..06a07052e 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e2m1_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e2m1_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e2m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e2m3_f32.ll index 741ae6b00..1ca414272 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e2m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e2m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e3m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e3m2_f32.ll index 8ccf28a96..ccb779f51 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e3m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e3m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e4m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e4m3_f32.ll index 5c195b249..c09c411b5 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e4m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e4m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e5m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e5m2_f32.ll index 7eb7c201a..4af500809 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e5m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e2m3_e5m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e2m1_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e2m1_f32.ll index 07e4ea706..a6f9c983c 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e2m1_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e2m1_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e2m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e2m3_f32.ll index f0b441420..366f21f1a 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e2m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e2m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e3m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e3m2_f32.ll index 203ff011b..07aad7bd2 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e3m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e3m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e4m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e4m3_f32.ll index 6eeeae590..026fd87d7 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e4m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e4m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e5m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e5m2_f32.ll index 40679ce7d..3696acfe1 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e5m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e3m2_e5m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e2m1_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e2m1_f32.ll index 7793fb1b6..7dd708b56 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e2m1_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e2m1_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e2m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e2m3_f32.ll index 186b31988..dde7c65b0 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e2m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e2m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e3m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e3m2_f32.ll index 4ecfffa2d..9c495a3a1 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e3m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e3m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e4m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e4m3_f32.ll index 00301b238..ed4c14092 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e4m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e4m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e5m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e5m2_f32.ll index 4584a6b8d..eb91c4228 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e5m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e4m3_e5m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e2m1_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e2m1_f32.ll index 36f7dab63..ebc6f8474 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e2m1_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e2m1_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e2m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e2m3_f32.ll index 5782ced89..a4c3b7880 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e2m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e2m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e3m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e3m2_f32.ll index 47e26c23f..16beafec7 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e3m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e3m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e4m3_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e4m3_f32.ll index 3be3033cc..6415441c0 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e4m3_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e4m3_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e5m2_f32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e5m2_f32.ll index 7fb9a3541..252ed1562 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e5m2_f32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_kind_f8f6f4_f32_e5m2_e5m2_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4.ll index b2ba7d385..20483e476 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_satfinite.ll index b06d99d19..e2cefbdfe 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_u4.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_u4.ll index 02c0aa515..d16920a06 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_u4_satfinite.ll index 9c406a2ca..355906bea 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s4_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8.ll index 64cf990e6..c44eab75b 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_satfinite.ll index 9c1a0875a..14ee6b4e0 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_u8.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_u8.ll index 22b414bd5..0e961fc73 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_u8_satfinite.ll index 063d77089..b85e8a31e 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_s8_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4.ll index d41ee4bd5..ed3fcdde2 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_s4.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_s4.ll index ff58561d3..44b49ffda 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_s4.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_s4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_s4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_s4_satfinite.ll index ea82f4255..895d4b076 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_s4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_s4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_satfinite.ll index 9a5019678..31ea21163 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u4_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8.ll index ad871e7a4..9e2876dda 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_s8.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_s8.ll index 1db5d5afc..4e9861522 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_s8.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_s8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_s8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_s8_satfinite.ll index 011ce74c4..7c47e2657 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_s8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_s8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_satfinite.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_satfinite.ll index 50ed30e6e..c3af92fbe 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_satfinite.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k64_s32_u8_satfinite.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k8_f32_tf32.ll b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k8_f32_tf32.ll index 292f2dced..7fa8cd9d1 100644 --- a/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k8_f32_tf32.ll +++ b/cuda-oxide/intrinsics/probes/mma_sp_ordered_metadata_m16n8k8_f32_tf32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/movmatrix_trans_b16.ll b/cuda-oxide/intrinsics/probes/movmatrix_trans_b16.ll index ac18f46e1..2d77d30f5 100644 --- a/cuda-oxide/intrinsics/probes/movmatrix_trans_b16.ll +++ b/cuda-oxide/intrinsics/probes/movmatrix_trans_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_bf16x2.ll b/cuda-oxide/intrinsics/probes/mul_bf16x2.ll index b02cc4168..94aa44491 100644 --- a/cuda-oxide/intrinsics/probes/mul_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/mul_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_f16x2.ll b/cuda-oxide/intrinsics/probes/mul_f16x2.ll index f29398d95..682f3ac26 100644 --- a/cuda-oxide/intrinsics/probes/mul_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/mul_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_f32x2.ll b/cuda-oxide/intrinsics/probes/mul_f32x2.ll index 0e826e31c..6686899b4 100644 --- a/cuda-oxide/intrinsics/probes/mul_f32x2.ll +++ b/cuda-oxide/intrinsics/probes/mul_f32x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_ftz_f32x2.ll b/cuda-oxide/intrinsics/probes/mul_ftz_f32x2.ll index 1dd4cb4bd..647cf62d7 100644 --- a/cuda-oxide/intrinsics/probes/mul_ftz_f32x2.ll +++ b/cuda-oxide/intrinsics/probes/mul_ftz_f32x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rm_f32.ll b/cuda-oxide/intrinsics/probes/mul_rm_f32.ll index 946d9ec9d..8f7c5b6b2 100644 --- a/cuda-oxide/intrinsics/probes/mul_rm_f32.ll +++ b/cuda-oxide/intrinsics/probes/mul_rm_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rm_f64.ll b/cuda-oxide/intrinsics/probes/mul_rm_f64.ll index 38869a4d3..705f6c6c7 100644 --- a/cuda-oxide/intrinsics/probes/mul_rm_f64.ll +++ b/cuda-oxide/intrinsics/probes/mul_rm_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rm_ftz_f32.ll b/cuda-oxide/intrinsics/probes/mul_rm_ftz_f32.ll index 930ba15be..d67454d0a 100644 --- a/cuda-oxide/intrinsics/probes/mul_rm_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/mul_rm_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rn_f32.ll b/cuda-oxide/intrinsics/probes/mul_rn_f32.ll index 8845444e2..afa88e6fd 100644 --- a/cuda-oxide/intrinsics/probes/mul_rn_f32.ll +++ b/cuda-oxide/intrinsics/probes/mul_rn_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rn_f64.ll b/cuda-oxide/intrinsics/probes/mul_rn_f64.ll index 9800e872e..95dbef9ae 100644 --- a/cuda-oxide/intrinsics/probes/mul_rn_f64.ll +++ b/cuda-oxide/intrinsics/probes/mul_rn_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rn_ftz_f32.ll b/cuda-oxide/intrinsics/probes/mul_rn_ftz_f32.ll index 1d01f663e..ac3828ba6 100644 --- a/cuda-oxide/intrinsics/probes/mul_rn_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/mul_rn_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rp_f32.ll b/cuda-oxide/intrinsics/probes/mul_rp_f32.ll index 6b5594af0..d46918c1c 100644 --- a/cuda-oxide/intrinsics/probes/mul_rp_f32.ll +++ b/cuda-oxide/intrinsics/probes/mul_rp_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rp_f64.ll b/cuda-oxide/intrinsics/probes/mul_rp_f64.ll index 68da7fde6..7f4d1da32 100644 --- a/cuda-oxide/intrinsics/probes/mul_rp_f64.ll +++ b/cuda-oxide/intrinsics/probes/mul_rp_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rp_ftz_f32.ll b/cuda-oxide/intrinsics/probes/mul_rp_ftz_f32.ll index 50d581c61..7c2f165e1 100644 --- a/cuda-oxide/intrinsics/probes/mul_rp_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/mul_rp_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rz_f32.ll b/cuda-oxide/intrinsics/probes/mul_rz_f32.ll index 75dd4ff99..87c5a6117 100644 --- a/cuda-oxide/intrinsics/probes/mul_rz_f32.ll +++ b/cuda-oxide/intrinsics/probes/mul_rz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rz_f64.ll b/cuda-oxide/intrinsics/probes/mul_rz_f64.ll index 2bfa2652f..a12e55757 100644 --- a/cuda-oxide/intrinsics/probes/mul_rz_f64.ll +++ b/cuda-oxide/intrinsics/probes/mul_rz_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/mul_rz_ftz_f32.ll b/cuda-oxide/intrinsics/probes/mul_rz_ftz_f32.ll index 5c3c0abd7..3caaa9fa8 100644 --- a/cuda-oxide/intrinsics/probes/mul_rz_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/mul_rz_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/nanosleep.ll b/cuda-oxide/intrinsics/probes/nanosleep.ll index 3d173156c..d7e614d79 100644 --- a/cuda-oxide/intrinsics/probes/nanosleep.ll +++ b/cuda-oxide/intrinsics/probes/nanosleep.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/neg_bf16x2.ll b/cuda-oxide/intrinsics/probes/neg_bf16x2.ll index e9488c24b..4ecbe2ba7 100644 --- a/cuda-oxide/intrinsics/probes/neg_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/neg_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/neg_f16x2.ll b/cuda-oxide/intrinsics/probes/neg_f16x2.ll index 385e6df41..7b748403e 100644 --- a/cuda-oxide/intrinsics/probes/neg_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/neg_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/nsmid.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/nsmid.lib_nvvm.ll index f9a042186..99a295448 100644 --- a/cuda-oxide/intrinsics/probes/nsmid.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/nsmid.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/nsmid.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/nsmid.llvm_nvptx.ll index fdda0fb0d..ecf9bcfc1 100644 --- a/cuda-oxide/intrinsics/probes/nsmid.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/nsmid.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/nwarpid.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/nwarpid.lib_nvvm.ll index 37efd6dc2..00e93ab2a 100644 --- a/cuda-oxide/intrinsics/probes/nwarpid.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/nwarpid.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/nwarpid.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/nwarpid.llvm_nvptx.ll index 83519dc2c..2ec773653 100644 --- a/cuda-oxide/intrinsics/probes/nwarpid.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/nwarpid.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/packed_atomic_add_bf16x2.ll b/cuda-oxide/intrinsics/probes/packed_atomic_add_bf16x2.ll index 4e227d273..ed0dea8e2 100644 --- a/cuda-oxide/intrinsics/probes/packed_atomic_add_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/packed_atomic_add_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/packed_atomic_add_f16x2.ll b/cuda-oxide/intrinsics/probes/packed_atomic_add_f16x2.ll index b31c4bb7f..c6589d012 100644 --- a/cuda-oxide/intrinsics/probes/packed_atomic_add_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/packed_atomic_add_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/pmevent.ll b/cuda-oxide/intrinsics/probes/pmevent.ll index 0fdf57f45..cee3b9742 100644 --- a/cuda-oxide/intrinsics/probes/pmevent.ll +++ b/cuda-oxide/intrinsics/probes/pmevent.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/prefetch_tma_descriptor.ll b/cuda-oxide/intrinsics/probes/prefetch_tma_descriptor.ll index f6ea49f51..69e33453c 100644 --- a/cuda-oxide/intrinsics/probes/prefetch_tma_descriptor.ll +++ b/cuda-oxide/intrinsics/probes/prefetch_tma_descriptor.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/prmt.ll b/cuda-oxide/intrinsics/probes/prmt.ll index 25e4ee79f..54e589832 100644 --- a/cuda-oxide/intrinsics/probes/prmt.ll +++ b/cuda-oxide/intrinsics/probes/prmt.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/prmt_b4e.ll b/cuda-oxide/intrinsics/probes/prmt_b4e.ll index 484a1dfdc..6b3114959 100644 --- a/cuda-oxide/intrinsics/probes/prmt_b4e.ll +++ b/cuda-oxide/intrinsics/probes/prmt_b4e.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/prmt_ecl.ll b/cuda-oxide/intrinsics/probes/prmt_ecl.ll index 34e0535dd..6455efc4b 100644 --- a/cuda-oxide/intrinsics/probes/prmt_ecl.ll +++ b/cuda-oxide/intrinsics/probes/prmt_ecl.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/prmt_ecr.ll b/cuda-oxide/intrinsics/probes/prmt_ecr.ll index f83f4a1c0..4128351cd 100644 --- a/cuda-oxide/intrinsics/probes/prmt_ecr.ll +++ b/cuda-oxide/intrinsics/probes/prmt_ecr.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/prmt_f4e.ll b/cuda-oxide/intrinsics/probes/prmt_f4e.ll index c4a761422..aae654597 100644 --- a/cuda-oxide/intrinsics/probes/prmt_f4e.ll +++ b/cuda-oxide/intrinsics/probes/prmt_f4e.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/prmt_rc16.ll b/cuda-oxide/intrinsics/probes/prmt_rc16.ll index 14192b0eb..7c2a19825 100644 --- a/cuda-oxide/intrinsics/probes/prmt_rc16.ll +++ b/cuda-oxide/intrinsics/probes/prmt_rc16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/prmt_rc8.ll b/cuda-oxide/intrinsics/probes/prmt_rc8.ll index 009328e41..a862ad045 100644 --- a/cuda-oxide/intrinsics/probes/prmt_rc8.ll +++ b/cuda-oxide/intrinsics/probes/prmt_rc8.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_approx_ftz_f32.ll b/cuda-oxide/intrinsics/probes/rcp_approx_ftz_f32.ll index c8ba34faf..04ce29599 100644 --- a/cuda-oxide/intrinsics/probes/rcp_approx_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_approx_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_approx_ftz_f64.ll b/cuda-oxide/intrinsics/probes/rcp_approx_ftz_f64.ll index e742043f1..e7fb891f2 100644 --- a/cuda-oxide/intrinsics/probes/rcp_approx_ftz_f64.ll +++ b/cuda-oxide/intrinsics/probes/rcp_approx_ftz_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rm_f32.ll b/cuda-oxide/intrinsics/probes/rcp_rm_f32.ll index 22099fbb9..7df0f7d12 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rm_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rm_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rm_f64.ll b/cuda-oxide/intrinsics/probes/rcp_rm_f64.ll index 9350e2b74..7c8b7d824 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rm_f64.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rm_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rm_ftz_f32.ll b/cuda-oxide/intrinsics/probes/rcp_rm_ftz_f32.ll index 366248017..63488a241 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rm_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rm_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rn_f32.ll b/cuda-oxide/intrinsics/probes/rcp_rn_f32.ll index 9b321528a..0a37391c9 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rn_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rn_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rn_f64.ll b/cuda-oxide/intrinsics/probes/rcp_rn_f64.ll index 4474b209b..ce085ea4b 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rn_f64.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rn_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rn_ftz_f32.ll b/cuda-oxide/intrinsics/probes/rcp_rn_ftz_f32.ll index 0706a5dd8..333cd323a 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rn_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rn_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rp_f32.ll b/cuda-oxide/intrinsics/probes/rcp_rp_f32.ll index ed85e8c9d..48ad43f70 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rp_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rp_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rp_f64.ll b/cuda-oxide/intrinsics/probes/rcp_rp_f64.ll index 67e94691f..0a5a25e5b 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rp_f64.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rp_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rp_ftz_f32.ll b/cuda-oxide/intrinsics/probes/rcp_rp_ftz_f32.ll index c55efd33b..1314397bb 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rp_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rp_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rz_f32.ll b/cuda-oxide/intrinsics/probes/rcp_rz_f32.ll index 8be6130b7..383459552 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rz_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rz_f64.ll b/cuda-oxide/intrinsics/probes/rcp_rz_f64.ll index 78a380cd8..2212561a3 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rz_f64.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rz_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rcp_rz_ftz_f32.ll b/cuda-oxide/intrinsics/probes/rcp_rz_ftz_f32.ll index 01d8028fd..08d83dfa2 100644 --- a/cuda-oxide/intrinsics/probes/rcp_rz_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/rcp_rz_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_add.ll b/cuda-oxide/intrinsics/probes/redux_sync_add.ll index 3c6fef28a..cd4bf7270 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_add.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_add.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_and.ll b/cuda-oxide/intrinsics/probes/redux_sync_and.ll index d62cd46fb..9664a17a4 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_and.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_and.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_max_abs_f32.ll b/cuda-oxide/intrinsics/probes/redux_sync_max_abs_f32.ll index 1977da8b3..f80cfa51c 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_max_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_max_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_max_abs_nan_f32.ll b/cuda-oxide/intrinsics/probes/redux_sync_max_abs_nan_f32.ll index 3da772f24..6773f431c 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_max_abs_nan_f32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_max_abs_nan_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_max_f32.ll b/cuda-oxide/intrinsics/probes/redux_sync_max_f32.ll index cd25209c1..ba1728e57 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_max_f32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_max_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_max_i32.ll b/cuda-oxide/intrinsics/probes/redux_sync_max_i32.ll index f25556e96..ec78783c2 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_max_i32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_max_i32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_max_nan_f32.ll b/cuda-oxide/intrinsics/probes/redux_sync_max_nan_f32.ll index cac9c75c9..82b368800 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_max_nan_f32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_max_nan_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_max_u32.ll b/cuda-oxide/intrinsics/probes/redux_sync_max_u32.ll index 6b3c0ab20..a4a13febf 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_max_u32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_max_u32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_min_abs_f32.ll b/cuda-oxide/intrinsics/probes/redux_sync_min_abs_f32.ll index 8fea47dc6..5f7dc786e 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_min_abs_f32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_min_abs_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_min_abs_nan_f32.ll b/cuda-oxide/intrinsics/probes/redux_sync_min_abs_nan_f32.ll index 6990a7d48..039a92320 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_min_abs_nan_f32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_min_abs_nan_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_min_f32.ll b/cuda-oxide/intrinsics/probes/redux_sync_min_f32.ll index 6de0160a5..28a7cad70 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_min_f32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_min_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_min_i32.ll b/cuda-oxide/intrinsics/probes/redux_sync_min_i32.ll index 5ef98bce2..3c06926ae 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_min_i32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_min_i32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_min_nan_f32.ll b/cuda-oxide/intrinsics/probes/redux_sync_min_nan_f32.ll index e375542b3..c98ffa085 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_min_nan_f32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_min_nan_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_min_u32.ll b/cuda-oxide/intrinsics/probes/redux_sync_min_u32.ll index 8b23ffc72..9362796a0 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_min_u32.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_min_u32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_or.ll b/cuda-oxide/intrinsics/probes/redux_sync_or.ll index 5def8ee0a..09ecebd52 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_or.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_or.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/redux_sync_xor.ll b/cuda-oxide/intrinsics/probes/redux_sync_xor.ll index ba934006d..06c05198b 100644 --- a/cuda-oxide/intrinsics/probes/redux_sync_xor.ll +++ b/cuda-oxide/intrinsics/probes/redux_sync_xor.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rsqrt_approx_f32.ll b/cuda-oxide/intrinsics/probes/rsqrt_approx_f32.ll index c3e17919e..8c6179c4a 100644 --- a/cuda-oxide/intrinsics/probes/rsqrt_approx_f32.ll +++ b/cuda-oxide/intrinsics/probes/rsqrt_approx_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rsqrt_approx_f64.ll b/cuda-oxide/intrinsics/probes/rsqrt_approx_f64.ll index 379c79050..312f889f5 100644 --- a/cuda-oxide/intrinsics/probes/rsqrt_approx_f64.ll +++ b/cuda-oxide/intrinsics/probes/rsqrt_approx_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/rsqrt_approx_ftz_f32.ll b/cuda-oxide/intrinsics/probes/rsqrt_approx_ftz_f32.ll index 345905cac..cfdf0a2b2 100644 --- a/cuda-oxide/intrinsics/probes/rsqrt_approx_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/rsqrt_approx_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/setmaxnreg_dec.ll b/cuda-oxide/intrinsics/probes/setmaxnreg_dec.ll index 3db958345..aa084717b 100644 --- a/cuda-oxide/intrinsics/probes/setmaxnreg_dec.ll +++ b/cuda-oxide/intrinsics/probes/setmaxnreg_dec.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/setmaxnreg_inc.ll b/cuda-oxide/intrinsics/probes/setmaxnreg_inc.ll index 5b81fdbd0..915ea2444 100644 --- a/cuda-oxide/intrinsics/probes/setmaxnreg_inc.ll +++ b/cuda-oxide/intrinsics/probes/setmaxnreg_inc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_down_f32_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_down_f32_sync.ll index 07931bc54..88f014a29 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_down_f32_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_down_f32_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_down_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_down_sync.ll index 565a566ec..92b62cccd 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_down_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_down_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_down_u64_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_down_u64_sync.ll index 9101ef9b0..6da995cc9 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_down_u64_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_down_u64_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_f32_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_f32_sync.ll index d08405a0c..75c48538f 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_f32_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_f32_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_sync.ll index c09516659..5c6fb324a 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_u64_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_u64_sync.ll index 1598ac517..20d248540 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_u64_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_u64_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_up_f32_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_up_f32_sync.ll index 968d5ae09..294c7862a 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_up_f32_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_up_f32_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_up_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_up_sync.ll index 9eab4166a..1a6568377 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_up_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_up_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_up_u64_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_up_u64_sync.ll index c3bdf2b0e..0d5545634 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_up_u64_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_up_u64_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_xor_f32_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_xor_f32_sync.ll index 8306292e9..2fb485631 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_xor_f32_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_xor_f32_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_xor_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_xor_sync.ll index f62c3be3f..41cef2d48 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_xor_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_xor_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/shuffle_xor_u64_sync.ll b/cuda-oxide/intrinsics/probes/shuffle_xor_u64_sync.ll index 34180ef98..5d5a590cf 100644 --- a/cuda-oxide/intrinsics/probes/shuffle_xor_u64_sync.ll +++ b/cuda-oxide/intrinsics/probes/shuffle_xor_u64_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sin_approx_f32.ll b/cuda-oxide/intrinsics/probes/sin_approx_f32.ll index bac49742c..77d2ba611 100644 --- a/cuda-oxide/intrinsics/probes/sin_approx_f32.ll +++ b/cuda-oxide/intrinsics/probes/sin_approx_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sin_approx_ftz_f32.ll b/cuda-oxide/intrinsics/probes/sin_approx_ftz_f32.ll index 4577db39c..ff161742e 100644 --- a/cuda-oxide/intrinsics/probes/sin_approx_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/sin_approx_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/smid.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/smid.lib_nvvm.ll index 1a59d8018..1a9c8758f 100644 --- a/cuda-oxide/intrinsics/probes/smid.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/smid.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/smid.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/smid.llvm_nvptx.ll index 286922323..50cce104a 100644 --- a/cuda-oxide/intrinsics/probes/smid.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/smid.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_approx_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_approx_f32.ll index 65b0fbe72..899e546d1 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_approx_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_approx_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_approx_ftz_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_approx_ftz_f32.ll index cd8eda50d..262b98ba4 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_approx_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_approx_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rm_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_rm_f32.ll index 2f866fb9a..d8928ed63 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rm_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rm_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rm_f64.ll b/cuda-oxide/intrinsics/probes/sqrt_rm_f64.ll index dd08cfa8d..da1f9136e 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rm_f64.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rm_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rm_ftz_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_rm_ftz_f32.ll index 9a52fb6c1..bfea216d8 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rm_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rm_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rn_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_rn_f32.ll index 6cb73537f..73b1e6071 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rn_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rn_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rn_f64.ll b/cuda-oxide/intrinsics/probes/sqrt_rn_f64.ll index 7f30e096c..50c4b939f 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rn_f64.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rn_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rn_ftz_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_rn_ftz_f32.ll index 111aff77c..c82e43546 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rn_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rn_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rp_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_rp_f32.ll index 7bf5b85fe..40f1721b6 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rp_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rp_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rp_f64.ll b/cuda-oxide/intrinsics/probes/sqrt_rp_f64.ll index fcd654a82..efc26e994 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rp_f64.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rp_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rp_ftz_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_rp_ftz_f32.ll index e0064049b..973d8f88c 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rp_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rp_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rz_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_rz_f32.ll index 161e597e6..db08de80b 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rz_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rz_f64.ll b/cuda-oxide/intrinsics/probes/sqrt_rz_f64.ll index 1ac36b7dc..1aba99e65 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rz_f64.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rz_f64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sqrt_rz_ftz_f32.ll b/cuda-oxide/intrinsics/probes/sqrt_rz_ftz_f32.ll index ed9a2b35c..dba46e92b 100644 --- a/cuda-oxide/intrinsics/probes/sqrt_rz_ftz_f32.ll +++ b/cuda-oxide/intrinsics/probes/sqrt_rz_ftz_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x2_b16.ll b/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x2_b16.ll index 23cecab80..16922e61a 100644 --- a/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x2_b16.ll +++ b/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x2_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x2_trans_b16.ll b/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x2_trans_b16.ll index ff62f080e..9a406fbf8 100644 --- a/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x2_trans_b16.ll +++ b/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x2_trans_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x4_b16.ll b/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x4_b16.ll index 7e927c768..208f8e3ba 100644 --- a/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x4_b16.ll +++ b/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x4_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x4_trans_b16.ll b/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x4_trans_b16.ll index 8dfad7bbd..1ea151eef 100644 --- a/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x4_trans_b16.ll +++ b/cuda-oxide/intrinsics/probes/stmatrix_m8n8_x4_trans_b16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sub_bf16x2.ll b/cuda-oxide/intrinsics/probes/sub_bf16x2.ll index fb4c2e2fe..7975adacf 100644 --- a/cuda-oxide/intrinsics/probes/sub_bf16x2.ll +++ b/cuda-oxide/intrinsics/probes/sub_bf16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sub_f16x2.ll b/cuda-oxide/intrinsics/probes/sub_f16x2.ll index f7ce7ff1c..9b500df35 100644 --- a/cuda-oxide/intrinsics/probes/sub_f16x2.ll +++ b/cuda-oxide/intrinsics/probes/sub_f16x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sub_f32x2.ll b/cuda-oxide/intrinsics/probes/sub_f32x2.ll index 182ae4c0b..a4e3cb2dd 100644 --- a/cuda-oxide/intrinsics/probes/sub_f32x2.ll +++ b/cuda-oxide/intrinsics/probes/sub_f32x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sub_ftz_f32x2.ll b/cuda-oxide/intrinsics/probes/sub_ftz_f32x2.ll index 0facbbe45..ba2a552a2 100644 --- a/cuda-oxide/intrinsics/probes/sub_ftz_f32x2.ll +++ b/cuda-oxide/intrinsics/probes/sub_ftz_f32x2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sync_mask.ll b/cuda-oxide/intrinsics/probes/sync_mask.ll index f6a9812a1..671f5e68b 100644 --- a/cuda-oxide/intrinsics/probes/sync_mask.ll +++ b/cuda-oxide/intrinsics/probes/sync_mask.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/sync_threads.ll b/cuda-oxide/intrinsics/probes/sync_threads.ll index 01a3bb1d2..23ac2c69d 100644 --- a/cuda-oxide/intrinsics/probes/sync_threads.ll +++ b/cuda-oxide/intrinsics/probes/sync_threads.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tanh_approx_f32.ll b/cuda-oxide/intrinsics/probes/tanh_approx_f32.ll index 725f001d1..4e0dad07a 100644 --- a/cuda-oxide/intrinsics/probes/tanh_approx_f32.ll +++ b/cuda-oxide/intrinsics/probes/tanh_approx_f32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_alloc.ll b/cuda-oxide/intrinsics/probes/tcgen05_alloc.ll index 4681ee92f..d8d494763 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_alloc.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_alloc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_alloc_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_alloc_cg2.ll index 188cef95f..d1b797d4b 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_alloc_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_alloc_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_commit.ll b/cuda-oxide/intrinsics/probes/tcgen05_commit.ll index b18d1467d..297994ef5 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_commit.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_commit.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_commit_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_commit_cg2.ll index de11e3ecb..5f335632a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_commit_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_commit_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_commit_multicast.ll b/cuda-oxide/intrinsics/probes/tcgen05_commit_multicast.ll index d6b22fd8a..680b6554e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_commit_multicast.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_commit_multicast.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_commit_multicast_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_commit_multicast_cg2.ll index eb3f9cc38..8186df1a2 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_commit_multicast_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_commit_multicast_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_commit_shared_cluster.ll b/cuda-oxide/intrinsics/probes/tcgen05_commit_shared_cluster.ll index 1db2dbdba..93669f83a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_commit_shared_cluster.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_commit_shared_cluster.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_commit_shared_cluster_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_commit_shared_cluster_cg2.ll index f5aa13b3e..c774b1911 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_commit_shared_cluster_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_commit_shared_cluster_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b.ll index 1e5fe4bb0..9d22837bc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b4x16_p64.ll index 19a692991..e829843a3 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b4x16_p64_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b4x16_p64_cg2.ll index c8d680b1c..be15604c5 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b4x16_p64_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b4x16_p64_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b6x16_p32.ll index 915f8808d..499bc4e5f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b6x16_p32_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b6x16_p32_cg2.ll index 272b8348b..21b479669 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b6x16_p32_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_b6x16_p32_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_cg2.ll index 934cd697e..604c8c017 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x128b_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b4x16_p64.ll index 3f3585ca3..46521f2e3 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b4x16_p64_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b4x16_p64_cg2.ll index bd9c2a8a6..4a3775a23 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b4x16_p64_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b4x16_p64_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b6x16_p32.ll index 6915054db..0734f6af1 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b6x16_p32_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b6x16_p32_cg2.ll index 2c93c4b8b..53b5263b4 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b6x16_p32_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_128x256b_b6x16_p32_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4.ll index 243e2f445..609b36a1a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b4x16_p64.ll index d7318b8ec..89fb8365b 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b4x16_p64_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b4x16_p64_cg2.ll index 49f8360d7..db65f0226 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b4x16_p64_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b4x16_p64_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b6x16_p32.ll index 1b0b9c842..46c530466 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b6x16_p32_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b6x16_p32_cg2.ll index fb9294fc0..1c488431c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b6x16_p32_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_b6x16_p32_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_cg2.ll index 1f7537312..5ebb0818a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_32x128b_warpx4_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b.ll index 705b185f6..8b3b1622a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b4x16_p64.ll index 5ad6aabca..314082c1f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b4x16_p64_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b4x16_p64_cg2.ll index 6123b0e8f..52df3ab37 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b4x16_p64_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b4x16_p64_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b6x16_p32.ll index 19e703d60..9107b0378 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b6x16_p32_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b6x16_p32_cg2.ll index 0a28e8367..98bd6d982 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b6x16_p32_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_b6x16_p32_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_cg2.ll index 8de4c9caf..bf2ee26c9 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_4x256b_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23.ll index 8b83fcc6e..5cf5b8076 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b4x16_p64.ll index 80745c11f..3fb56b30a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b4x16_p64_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b4x16_p64_cg2.ll index e224d3f2f..87a7dce95 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b4x16_p64_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b4x16_p64_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b6x16_p32.ll index b5fad6438..3af2fa6e6 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b6x16_p32_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b6x16_p32_cg2.ll index b4803dae6..b2219aa9a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b6x16_p32_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_b6x16_p32_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_cg2.ll index 9d5209486..98558fc3b 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_01_23_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13.ll index dab75d57b..1f9def3fe 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b4x16_p64.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b4x16_p64.ll index 4ada1ddd7..fb2acc968 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b4x16_p64.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b4x16_p64.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b4x16_p64_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b4x16_p64_cg2.ll index 4e70b6a4c..d98278527 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b4x16_p64_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b4x16_p64_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b6x16_p32.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b6x16_p32.ll index 1033e04fb..3fc401988 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b6x16_p32.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b6x16_p32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b6x16_p32_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b6x16_p32_cg2.ll index 916530764..f7547ba80 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b6x16_p32_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_b6x16_p32_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_cg2.ll index 2645ba6aa..27b05800c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_64x128b_warpx2_02_13_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_smem_to_tmem.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_smem_to_tmem.ll index 0325551ea..df8ca1696 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_smem_to_tmem.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_smem_to_tmem.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_cp_smem_to_tmem_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_cp_smem_to_tmem_cg2.ll index 493cc35b9..511b132a1 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_cp_smem_to_tmem_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_cp_smem_to_tmem_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_dealloc.ll b/cuda-oxide/intrinsics/probes/tcgen05_dealloc.ll index fc9d5dd47..3951b4de4 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_dealloc.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_dealloc.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_dealloc_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_dealloc_cg2.ll index 834b5cd91..4df5378fe 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_dealloc_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_dealloc_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_fence_after_thread_sync.ll b/cuda-oxide/intrinsics/probes/tcgen05_fence_after_thread_sync.ll index badff3493..26cf54f7b 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_fence_after_thread_sync.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_fence_after_thread_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_fence_before_thread_sync.ll b/cuda-oxide/intrinsics/probes/tcgen05_fence_before_thread_sync.ll index da4f8c90e..3a19ad72e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_fence_before_thread_sync.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_fence_before_thread_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x16_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x16_pack16.ll index 510ca76c4..1ec0cdcaa 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x16_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x16_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x16_raw.ll index 39793770f..c88ad5184 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x1_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x1_pack16.ll index 6558b2993..da5e04291 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x1_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x1_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x1_raw.ll index 5943a61b0..6303933af 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x2_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x2_pack16.ll index 2c617ef28..42b400045 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x2_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x2_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x2_raw.ll index 3f16f4c1b..9cd59f584 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x32_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x32_pack16.ll index dabd22d99..b7d8da54e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x32_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x32_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x32_raw.ll index 9a2103496..55526b409 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x4_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x4_pack16.ll index 5404838fb..67f98f937 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x4_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x4_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x4_raw.ll index e4e364709..3ab5f9d7f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x64_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x64_pack16.ll index 34e7e8520..1b9781971 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x64_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x64_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x64_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x64_raw.ll index d10bdcb5c..244065eed 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x64_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x64_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x8_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x8_pack16.ll index 49f9fad5c..5194ca736 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x8_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x8_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x8_raw.ll index 259a47bab..7760557e8 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x128b_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_pure.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_pure.ll index ecc652aaa..54d00ae6f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_pure.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_pure.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x16_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x16_pack16.ll index fedec7933..85d311b1f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x16_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x16_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x16_raw.ll index 501dac068..51a1cc4be 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x1_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x1_pack16.ll index c81c06e7b..e2ce151b3 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x1_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x1_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x1_raw.ll index c77bd824d..1d73cdf17 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x2_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x2_pack16.ll index ac2162f91..e27fb796c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x2_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x2_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x2_raw.ll index 8ba23ed5a..4a9a9370d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x32_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x32_pack16.ll index 6e1f95578..ac05fcaf8 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x32_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x32_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x32_raw.ll index bb00237f0..f7af082e9 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x4_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x4_pack16.ll index 68482cc33..6546558e0 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x4_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x4_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x4_raw.ll index 59defae2a..20f21f5a9 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_pack16.ll index bef54212c..72b713090 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_pure.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_pure.ll index 8f5b6ac36..6150ff47a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_pure.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_pure.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_raw.ll index 070cbbc91..a266721e6 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x256b_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x128_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x128_pack16.ll index ad594f196..cfa637772 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x128_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x128_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x128_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x128_raw.ll index 60d37021f..a353a9b6f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x128_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x128_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x16_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x16_pack16.ll index c57c9da0a..d215a0615 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x16_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x16_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x16_raw.ll index 4d818e944..98494570f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x1_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x1_pack16.ll index ece544659..1ce69e164 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x1_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x1_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x1_raw.ll index 376b7b810..3eeac417c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x2_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x2_pack16.ll index 4fdc4403e..6382b60a7 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x2_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x2_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x2_raw.ll index 680541778..72ebd6e3d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x32_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x32_pack16.ll index be22d6b1f..0389053fc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x32_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x32_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x32_raw.ll index 6a1ea536a..68220bafb 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x4_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x4_pack16.ll index b584ba9d9..8cddcb27f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x4_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x4_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x4_raw.ll index 34458c188..fc10354b3 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x64_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x64_pack16.ll index 3ca138300..19337137c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x64_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x64_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x64_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x64_raw.ll index 0dad60185..58793d916 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x64_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x64_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x8_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x8_pack16.ll index eea053e0e..89a1b7d19 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x8_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x8_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x8_raw.ll index e56db75f7..cd85db48a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x32bx2_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x128_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x128_pack16.ll index f54a4b971..bfe73d3fd 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x128_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x128_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x128_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x128_raw.ll index 662a79b2a..758de0cd9 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x128_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x128_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x16_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x16_pack16.ll index 3015ca305..88d5b6978 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x16_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x16_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x16_raw.ll index 5c9a851cf..d0a4d2bfd 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x1_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x1_pack16.ll index 548524be6..241f31846 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x1_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x1_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x1_raw.ll index ee8e28872..d36eb8bc4 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x2_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x2_pack16.ll index 6f7edcec9..8eb4a6807 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x2_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x2_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x2_raw.ll index 5098ca6a0..964c7a0ea 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x32_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x32_pack16.ll index 9236a7e66..06b722500 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x32_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x32_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x32_raw.ll index 9e8ccc97e..722c1e76e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x4_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x4_pack16.ll index c3dc0adef..99de889a0 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x4_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x4_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x4_raw.ll index 5630c662c..987c80b11 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x64_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x64_pack16.ll index 18bf3328b..469c64b5c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x64_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x64_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x64_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x64_raw.ll index da5a8ac72..9cbb7e9dc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x64_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x64_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x8_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x8_pack16.ll index ea903e9e9..8a64fef30 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x8_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x8_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x8_raw.ll index 3ce9fe0bb..6799a8b98 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_16x64b_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x128_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x128_pack16.ll index 7f5c30cfa..4fefdf518 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x128_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x128_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x128_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x128_raw.ll index 6a9ba29f7..55eab790c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x128_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x128_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x16_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x16_pack16.ll index c86e7781f..de3a3688d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x16_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x16_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x16_raw.ll index b90be4e66..48b529759 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x1_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x1_pack16.ll index 15b87faad..a34f46a99 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x1_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x1_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x1_raw.ll index 33344e3c0..a297012fc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x2_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x2_pack16.ll index b17d114dc..0b77d9014 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x2_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x2_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x2_raw.ll index af1d53993..a4cbcd4a7 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x32_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x32_pack16.ll index 8bee1d37a..5c6c12b1c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x32_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x32_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x32_raw.ll index 2c05ecb2c..7f7c2d63c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x4_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x4_pack16.ll index 0e80981b8..34cf97ee5 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x4_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x4_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x4_raw.ll index fa8fdfa09..9c3dd0793 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x64_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x64_pack16.ll index c3cbb87b0..bb372c37c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x64_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x64_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x64_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x64_raw.ll index ea36ca43d..beae73def 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x64_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x64_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x8_pack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x8_pack16.ll index 6216ada9a..409acdcb0 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x8_pack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x8_pack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x8_raw.ll index 8da86c672..74ad0bd0e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_ld_32x32b_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_load_wait.ll b/cuda-oxide/intrinsics/probes/tcgen05_load_wait.ll index 595b69fa2..3deb9f256 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_load_wait.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_load_wait.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_e2m1.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_e2m1.ll index 573fed3b4..42800953d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_e2m3.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_e2m3.ll index 27262fdf3..849a8ced5 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_e3m2.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_e3m2.ll index 6a61de230..310f7492e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_e4m3.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_e4m3.ll index 98d37ad50..1a454371b 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_e5m2.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_e5m2.ll index 6c40df6ff..26fb8d385 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_f16.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_f16.ll index 5308f4313..29c6f223a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_f16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_f16_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_f16_cg2.ll index 47f8ca3dd..98de1423c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_f16_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_f16_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_shared.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_shared.ll index 4cd8bba4a..27ee81805 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_shared.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_shared.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_shared.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_shared.ll index 8bb03d2af..1e0cc19dd 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_shared.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_shared.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_tensor.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_tensor.ll index bb1f77625..6abcd38f6 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_tensor.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_tensor.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_tensor_ashift.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_tensor_ashift.ll index fbd424203..9195bdca3 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_tensor_ashift.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_sp_tensor_ashift.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_tensor.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_tensor.ll index 9dbba4242..eff22a182 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_tensor.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_tensor.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_tensor_ashift.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_tensor_ashift.ll index 9c2515b6a..ac63ae7be 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_tensor_ashift.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_tensor_ashift.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_bf16.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_bf16.ll index 8db4ffbae..c7d14dd9f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_bf16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_bf16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e2m1.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e2m1.ll index c362ea1a3..96e1adbd6 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e2m1.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e2m1.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e2m3.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e2m3.ll index b92b3aec0..5031449cd 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e2m3.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e2m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e3m2.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e3m2.ll index 042ce7ad7..fa0529a64 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e3m2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e3m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e4m3.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e4m3.ll index 9a620b830..09d67dc49 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e4m3.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e4m3.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e5m2.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e5m2.ll index 9f3ce75d0..752410fb1 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e5m2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_e5m2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_f16.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_f16.ll index e0e6c7c07..0a13bb8f8 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_f16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_f16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_shared.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_shared.ll index 09ea38c76..5743c9c4c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_shared.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_shared.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_shared_zero_col_mask.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_shared_zero_col_mask.ll index ed121fd4c..07d00be37 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_shared_zero_col_mask.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_shared_zero_col_mask.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_shared.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_shared.ll index 9e0601053..e43283a93 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_shared.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_shared.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_shared_zero_col_mask.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_shared_zero_col_mask.ll index 1caaecb0f..7b2b82371 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_shared_zero_col_mask.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_shared_zero_col_mask.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_tensor.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_tensor.ll index 28ac79fae..f6218e95e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_tensor.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_tensor.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_tensor_zero_col_mask.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_tensor_zero_col_mask.ll index e50233867..35ad46af8 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_tensor_zero_col_mask.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_sp_tensor_zero_col_mask.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tensor.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tensor.ll index 3906ca1f4..b9863df9f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tensor.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tensor.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tensor_zero_col_mask.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tensor_zero_col_mask.ll index 3289dc793..22d912ebc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tensor_zero_col_mask.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tensor_zero_col_mask.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tf32.ll b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tf32.ll index cae6ddcfb..6a5f3f71d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tf32.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_mma_ws_tf32.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_relinquish_alloc_permit.ll b/cuda-oxide/intrinsics/probes/tcgen05_relinquish_alloc_permit.ll index fbad87955..b9bfa4e28 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_relinquish_alloc_permit.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_relinquish_alloc_permit.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_relinquish_alloc_permit_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_relinquish_alloc_permit_cg2.ll index ddfcc6494..b476ca18b 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_relinquish_alloc_permit_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_relinquish_alloc_permit_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_shift_down.ll b/cuda-oxide/intrinsics/probes/tcgen05_shift_down.ll index 508610005..709fd4c78 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_shift_down.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_shift_down.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_shift_down_cg2.ll b/cuda-oxide/intrinsics/probes/tcgen05_shift_down_cg2.ll index 400398773..cffdee798 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_shift_down_cg2.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_shift_down_cg2.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x16_raw.ll index 1b81acc23..96da6e4d7 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x16_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x16_unpack16.ll index 0d35a68a1..0a275483e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x16_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x16_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x1_raw.ll index aa5dcdad3..903175377 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x1_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x1_unpack16.ll index 24ba3b65f..520c730fa 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x1_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x1_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x2_raw.ll index 7b375acf9..75ba7b4dc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x2_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x2_unpack16.ll index 1347bdbd3..2149154a6 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x2_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x2_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x32_raw.ll index 8ce48b98b..88f960628 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x32_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x32_unpack16.ll index 7c707115e..df77e3971 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x32_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x32_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x4_raw.ll index 29fd5f821..f24fa552a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x4_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x4_unpack16.ll index 24a83eeca..d99ccc339 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x4_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x4_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x64_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x64_raw.ll index 7e08f3cc6..8ee1f92bf 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x64_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x64_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x64_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x64_unpack16.ll index 7359c1581..4a0443340 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x64_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x64_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x8_raw.ll index 7508d9512..a9dab820b 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x8_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x8_unpack16.ll index 81c394680..5ec0e410d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x8_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x128b_x8_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x16_raw.ll index dd617432b..f7ce12dcd 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x16_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x16_unpack16.ll index b1431a608..908f347fc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x16_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x16_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x1_raw.ll index dfd391a87..4efbb6d01 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x1_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x1_unpack16.ll index 113cd6f2c..8a693b260 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x1_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x1_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x2_raw.ll index 65278745c..9003d5c6f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x2_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x2_unpack16.ll index 32219dc01..cf5091124 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x2_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x2_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x32_raw.ll index c2616feaa..fe984bd8d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x32_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x32_unpack16.ll index bab355423..3f54d5f4f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x32_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x32_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x4_raw.ll index ef892661c..4618b8450 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x4_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x4_unpack16.ll index 1d16dd79e..040c863cc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x4_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x4_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x8_raw.ll index b3c367164..a867eeb9a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x8_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x8_unpack16.ll index 1649c07be..1600215ba 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x8_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x256b_x8_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x128_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x128_raw.ll index ea6a58618..b11371319 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x128_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x128_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x128_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x128_unpack16.ll index 5fc9cc0c3..5b4a21529 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x128_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x128_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x16_raw.ll index 76c9b5e12..7e07817a7 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x16_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x16_unpack16.ll index 281a56d4e..63b694a8e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x16_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x16_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x1_raw.ll index 2fa294582..f192065fb 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x1_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x1_unpack16.ll index d0d8759e9..c466d4ba2 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x1_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x1_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x2_raw.ll index 6710bd6d4..6b35b0ac5 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x2_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x2_unpack16.ll index fa54eb172..e2aa2f966 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x2_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x2_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x32_raw.ll index ddda0f9b8..2595bb91e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x32_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x32_unpack16.ll index 9989a0013..0bd7fc4c5 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x32_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x32_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x4_raw.ll index 5cfb0ea15..39f313995 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x4_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x4_unpack16.ll index da536ae70..296df3ecd 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x4_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x4_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x64_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x64_raw.ll index dc58a9dc2..2fae59724 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x64_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x64_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x64_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x64_unpack16.ll index 3af785f0b..584ee36db 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x64_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x64_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x8_raw.ll index 68bd22e8f..f6471e38c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x8_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x8_unpack16.ll index 7b36c3f64..4f3563faa 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x8_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x32bx2_x8_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x128_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x128_raw.ll index eea2ea995..3dc41182f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x128_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x128_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x128_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x128_unpack16.ll index ac3c80e5a..f245938c2 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x128_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x128_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x16_raw.ll index 52564ec66..07db680da 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x16_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x16_unpack16.ll index 2a3c96178..66e6f15c9 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x16_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x16_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x1_raw.ll index cae93a7a6..ec1f06584 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x1_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x1_unpack16.ll index 506ea6257..0dd568d55 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x1_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x1_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x2_raw.ll index 2dddd7ce9..fc9886ad4 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x2_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x2_unpack16.ll index 3994c8a4b..84e29c19f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x2_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x2_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x32_raw.ll index 15fb8957f..78232c22b 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x32_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x32_unpack16.ll index 5d05ee095..d678d472c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x32_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x32_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x4_raw.ll index f2fae2861..9aee0baaf 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x4_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x4_unpack16.ll index 73289f854..928d01c2d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x4_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x4_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x64_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x64_raw.ll index acc70601f..12a6c4e9d 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x64_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x64_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x64_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x64_unpack16.ll index 2574645ac..2c528316f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x64_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x64_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x8_raw.ll index cf8ac5941..399c847dc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x8_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x8_unpack16.ll index f7fdb470e..a98fc65a8 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x8_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_16x64b_x8_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x128_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x128_raw.ll index b2b72b7be..dfe8fb10a 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x128_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x128_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x128_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x128_unpack16.ll index 51c111b66..18dc36021 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x128_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x128_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x16_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x16_raw.ll index e905a65cc..748b1f9b3 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x16_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x16_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x16_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x16_unpack16.ll index dfce971db..943c7cbec 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x16_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x16_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x1_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x1_raw.ll index a0daa1fa6..5f043c91c 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x1_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x1_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x1_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x1_unpack16.ll index 27f2dadbc..7c7301e2f 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x1_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x1_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x2_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x2_raw.ll index 65292054b..375ddcae9 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x2_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x2_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x2_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x2_unpack16.ll index 8dc86b0fb..7c0fad985 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x2_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x2_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x32_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x32_raw.ll index 04be3d6a6..c7b71fa69 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x32_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x32_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x32_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x32_unpack16.ll index 0ce49b786..0fbd11caf 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x32_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x32_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x4_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x4_raw.ll index 3bd1c7461..816202a82 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x4_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x4_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x4_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x4_unpack16.ll index 9e8e75fda..245dc44bc 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x4_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x4_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x64_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x64_raw.ll index 269ef0773..9ec39e8d5 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x64_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x64_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x64_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x64_unpack16.ll index d24ba123b..b29585282 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x64_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x64_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x8_raw.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x8_raw.ll index 25074b9b4..3f9ab7e0e 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x8_raw.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x8_raw.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x8_unpack16.ll b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x8_unpack16.ll index 423a7b40f..0d02551a7 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x8_unpack16.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_st_32x32b_x8_unpack16.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tcgen05_store_wait.ll b/cuda-oxide/intrinsics/probes/tcgen05_store_wait.ll index a90834a82..4bcbb78e4 100644 --- a/cuda-oxide/intrinsics/probes/tcgen05_store_wait.ll +++ b/cuda-oxide/intrinsics/probes/tcgen05_store_wait.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_box_dim.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_box_dim.ll index 0b8a01a19..e570b7310 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_box_dim.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_box_dim.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_element_stride.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_element_stride.ll index 576978840..3e3826968 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_element_stride.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_element_stride.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_element_type.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_element_type.ll index 786ce3e14..01f5d53a2 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_element_type.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_element_type.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_fill_mode.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_fill_mode.ll index 983a5c7f4..4ebefec49 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_fill_mode.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_fill_mode.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_global_address.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_global_address.ll index 6be7af57f..c3dfc1223 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_global_address.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_global_address.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_global_dim.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_global_dim.ll index 69f632e4f..046affc44 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_global_dim.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_global_dim.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_global_stride.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_global_stride.ll index c3528e6c9..ff3b827b2 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_global_stride.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_global_stride.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_interleave_layout.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_interleave_layout.ll index dcf8a4502..31b193d50 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_interleave_layout.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_interleave_layout.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_rank.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_rank.ll index b82acad78..64984bfa3 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_rank.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_rank.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_swizzle_atomicity.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_swizzle_atomicity.ll index e7ca4939e..8b85971e9 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_swizzle_atomicity.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_swizzle_atomicity.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/tensormap_replace_swizzle_mode.ll b/cuda-oxide/intrinsics/probes/tensormap_replace_swizzle_mode.ll index f0faaacf6..79b2160a5 100644 --- a/cuda-oxide/intrinsics/probes/tensormap_replace_swizzle_mode.ll +++ b/cuda-oxide/intrinsics/probes/tensormap_replace_swizzle_mode.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/thread_idx_x.ll b/cuda-oxide/intrinsics/probes/thread_idx_x.ll index 861f7b549..4d8cb8289 100644 --- a/cuda-oxide/intrinsics/probes/thread_idx_x.ll +++ b/cuda-oxide/intrinsics/probes/thread_idx_x.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/thread_idx_y.ll b/cuda-oxide/intrinsics/probes/thread_idx_y.ll index b9c97e960..17e9c6813 100644 --- a/cuda-oxide/intrinsics/probes/thread_idx_y.ll +++ b/cuda-oxide/intrinsics/probes/thread_idx_y.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/thread_idx_z.ll b/cuda-oxide/intrinsics/probes/thread_idx_z.ll index 074d5c662..78dfc2d2b 100644 --- a/cuda-oxide/intrinsics/probes/thread_idx_z.ll +++ b/cuda-oxide/intrinsics/probes/thread_idx_z.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/threadfence.ll b/cuda-oxide/intrinsics/probes/threadfence.ll index 81bb0b1da..0b8fc8eb4 100644 --- a/cuda-oxide/intrinsics/probes/threadfence.ll +++ b/cuda-oxide/intrinsics/probes/threadfence.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/threadfence_block.ll b/cuda-oxide/intrinsics/probes/threadfence_block.ll index 2422f916a..50af1914b 100644 --- a/cuda-oxide/intrinsics/probes/threadfence_block.ll +++ b/cuda-oxide/intrinsics/probes/threadfence_block.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/threadfence_system.ll b/cuda-oxide/intrinsics/probes/threadfence_system.ll index d5181821c..8afc25bda 100644 --- a/cuda-oxide/intrinsics/probes/threadfence_system.ll +++ b/cuda-oxide/intrinsics/probes/threadfence_system.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/total_smem_size.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/total_smem_size.lib_nvvm.ll index 135eb1131..ccee72572 100644 --- a/cuda-oxide/intrinsics/probes/total_smem_size.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/total_smem_size.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/total_smem_size.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/total_smem_size.llvm_nvptx.ll index 9124402f8..9387c0646 100644 --- a/cuda-oxide/intrinsics/probes/total_smem_size.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/total_smem_size.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/trap.ll b/cuda-oxide/intrinsics/probes/trap.ll index 91eec4359..23efc3bf9 100644 --- a/cuda-oxide/intrinsics/probes/trap.ll +++ b/cuda-oxide/intrinsics/probes/trap.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/uni_sync.ll b/cuda-oxide/intrinsics/probes/uni_sync.ll index b990da9cc..5987a4d7c 100644 --- a/cuda-oxide/intrinsics/probes/uni_sync.ll +++ b/cuda-oxide/intrinsics/probes/uni_sync.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/warpid.lib_nvvm.ll b/cuda-oxide/intrinsics/probes/warpid.lib_nvvm.ll index b36a3105c..2d1ddd17b 100644 --- a/cuda-oxide/intrinsics/probes/warpid.lib_nvvm.ll +++ b/cuda-oxide/intrinsics/probes/warpid.lib_nvvm.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/warpid.llvm_nvptx.ll b/cuda-oxide/intrinsics/probes/warpid.llvm_nvptx.ll index c74745965..16fea2fef 100644 --- a/cuda-oxide/intrinsics/probes/warpid.llvm_nvptx.ll +++ b/cuda-oxide/intrinsics/probes/warpid.llvm_nvptx.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/wgmma_commit_group.ll b/cuda-oxide/intrinsics/probes/wgmma_commit_group.ll index 3584d85f0..52ef66f4d 100644 --- a/cuda-oxide/intrinsics/probes/wgmma_commit_group.ll +++ b/cuda-oxide/intrinsics/probes/wgmma_commit_group.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/wgmma_fence.ll b/cuda-oxide/intrinsics/probes/wgmma_fence.ll index a2b40a4c8..6744b757f 100644 --- a/cuda-oxide/intrinsics/probes/wgmma_fence.ll +++ b/cuda-oxide/intrinsics/probes/wgmma_fence.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/intrinsics/probes/wgmma_wait_group.ll b/cuda-oxide/intrinsics/probes/wgmma_wait_group.ll index 8ba5bde0b..e269d7ec9 100644 --- a/cuda-oxide/intrinsics/probes/wgmma_wait_group.ll +++ b/cuda-oxide/intrinsics/probes/wgmma_wait_group.ll @@ -1,7 +1,7 @@ ; SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. ; SPDX-License-Identifier: Apache-2.0 ; @generated by cuda-intrinsics-gen 0.1.0. DO NOT EDIT. -; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 31b5238818203ac4c636984d416ea7791c5861feb0ad30d1f5fc693ebefb4399. +; Catalog schema/version: 46/0.1.0; intrinsic ABI: v1; catalog SHA-256: 7e6b4b4b3b5958a0a36fb0c40486b8263e0799fbe1f23cb9be64b6687f27e3a0. ; LLVM source: aab11ee0ea5ea9a1d462ad0bc1fae9d9bb643019. target triple = "nvptx64-nvidia-cuda" diff --git a/cuda-oxide/scripts/smoketest.sh b/cuda-oxide/scripts/smoketest.sh index 0753ef779..35d8eb57a 100755 --- a/cuda-oxide/scripts/smoketest.sh +++ b/cuda-oxide/scripts/smoketest.sh @@ -97,8 +97,8 @@ AUTO_NVVM_EXAMPLES=(libdevice_math) IKET_EXAMPLES=(iket_trace) BLACKWELL_COMPILE_EXAMPLES=(generated_intrinsics_blackwell) SM100_COMPILE_EXAMPLES=(redux_f32) -NVVM_VERIFY_EXAMPLES=(constant_memory cp_async_small device_global enum_constant_provenance ex2_approx_f16 generated_intrinsics generated_intrinsics_blackwell generated_ldmatrix kernel_reference_validity legacy_atomic_fadd legacy_atomic_rmw_cas libdevice_math legacy_nvvm_pointer_shapes packed_atomic_add primitive_stress scoped_atomic_load_store shuffle_64 tcgen05 tuple_constant_provenance wgmma_mma_bf16) -ERROR_EXAMPLES=(error error_set_discriminant_uninhabited error_enum_bool_payload_addr error_enum_pointer_overlap error_enum_shared_pointer_layout error_heap_alloc error_host_arch_intrinsic error_host_target_feature error_kernel_shared_param error_missing_device_attr error_generated_intrinsic_abi error_generated_intrinsic_unknown_id error_generated_intrinsic_fn_pointer error_generated_intrinsic_callable error_union_shared_pointer_addr error_union_shared_pointer_layout) +NVVM_VERIFY_EXAMPLES=(constant_memory cp_async_small device_global enum_constant_provenance ex2_approx_f16 generated_intrinsics generated_intrinsics_blackwell generated_ldmatrix kernel_reference_validity legacy_atomic_fadd legacy_atomic_rmw_cas libdevice_math legacy_nvvm_pointer_shapes packed_atomic_add primitive_stress scoped_atomic_load_store shared_ptr_api shuffle_64 tcgen05 tuple_constant_provenance wgmma_mma_bf16) +ERROR_EXAMPLES=(error error_set_discriminant_uninhabited error_enum_bool_payload_addr error_enum_pointer_overlap error_enum_shared_pointer_layout error_heap_alloc error_host_arch_intrinsic error_host_target_feature error_kernel_shared_param error_kernel_shared_ptr_param error_missing_device_attr error_generated_intrinsic_abi error_generated_intrinsic_unknown_id error_generated_intrinsic_fn_pointer error_generated_intrinsic_callable error_union_shared_pointer_addr error_union_shared_pointer_layout) # Per-example rustc flags policy: an example that needs a special rustc flag # (e.g. disjoint_slice_len needs -Zinline-mir=no to keep its regression @@ -125,7 +125,7 @@ NO_LAUNCH_EXAMPLES=(wgmma_mma_bf16) # emitted before opt folds and reorders it. The optimized build still runs and # still decides the verdict; these get one extra CUDA_OXIDE_NO_OPT=1 build # afterwards purely to produce the artifacts the shape check reads. -NO_OPT_SHAPE_EXAMPLES=(const_bool_dead_branch) +NO_OPT_SHAPE_EXAMPLES=(const_bool_dead_branch shared_memory_helper) # This example intentionally configures a cubin artifact in Cargo metadata. # Its supported project route necessarily includes NVVM IR + libNVVM + @@ -702,6 +702,15 @@ verdict_error() { return 1 fi ;; + error_kernel_shared_ptr_param) + # The nested form's marker differs from the bare parameter's + # `is a pointer into shared memory`, so this fixture proves the + # exporter looked inside the parameter struct. + if ! grep -Fq 'holds a pointer into shared memory' "${log}"; then + echo "FAIL (missing nested shared-memory kernel-parameter diagnostic)" + return 1 + fi + ;; error_generated_intrinsic_abi) if ! grep -Fq 'cuda-intrinsics ABI mismatch' "${log}" \ || ! grep -Fq '__cuda_oxide_intrinsic_abi_v2::i0001' "${log}"; then