llama.cpp
ggml-cpu: enable tiled flash attention for non-vector-multiple head dims on x86
#29423
Open
Open
ggml-cpu: enable tiled flash attention for non-vector-multiple head dims on x86 #29423
SongXiaoXi
wants to merge
2 commits into
ggml-org:master
from
SongXiaoXi:cpu_fa_tile
ggml-cpu: enable tiled flash attention for non-vector-multiple head d…
4d5e0eac
am17an
commented
on 2026-09-25
add AVX2 support for masked loading and storing in simd_gemm_ukernel_…
9257ffe5
Login to write a write a comment.
Login via GitHub