llama: GDN rollback by replay, one-graph MTP draft chain, K/V-only MTP prompt pass, sync trim (port of Torchit1/arc-b580) - #358

llama: GDN rollback by replay, one-graph MTP draft chain, K/V-only MTP prompt pass, sync trim (port of Torchit1/arc-b580)#358

bri-prism merged 66 commits into

prismPrismML-Eng/llama.cpp:prismfrom

sycl/arc-b580-port-llamaPrismML-Eng/llama.cpp:sycl/arc-b580-port-llamaCopy head branch name to clipboard