Skip to content

Releases: iacopPBK/llama.cpp-gfx906

b6637

29 Sep 17:49

Choose a tag to compare

Optimize AMD GFX906 flash attention with DS_SWIZZLE instrinsics Autho…

GFX906 Kernels Optimizations v0.0.2

27 Sep 13:58

Choose a tag to compare

Hard swizzling to properly exploit gfx906 waves reduction

v0.0.1 GFX906 Kernels Optimizations v0.0.1

08 Sep 15:12

Choose a tag to compare

First release for optimizing the llama cpp on our beloved cheap and slow videocards.