Repository navigation

ggml-cpu : AVX-512 tiled kernel for single-token flash attention (faster long-context decode) - #254

Merged

bri-prism merged 1 commit into

PrismML-Eng:prismfrom

lenny76:cpu-fa-decode-avx512

Sep 25, 2026

Merged

ggml-cpu : AVX-512 tiled kernel for single-token flash attention (faster long-context decode)#254
bri-prism merged 1 commit into
PrismML-Eng:prismfrom
lenny76:cpu-fa-decode-avx512

Commits on Sep 23, 2026
- ggml-cpu : AVX-512 tiled kernel for single-token flash attention
  
  lenny76
  and
  claude
  committed
  5e95a48