]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
vulkan: use vec dot for matrix matrix multiplications (llama/16056)
authorRuben Ortlam <redacted>
Sat, 20 Sep 2025 08:42:56 +0000 (10:42 +0200)
committerGeorgi Gerganov <redacted>
Sat, 20 Sep 2025 10:33:50 +0000 (13:33 +0300)
commit8de5c8236009f17348398117f38f8a95a2232e77
treebcb4b520f730b50807dcd722b2cb8e823a806393
parentd98d650b626c070698c70fec588a9d7506a7538a
vulkan: use vec dot for matrix matrix multiplications (llama/16056)

* vulkan: Change the mul_mm shared memory and register caching system to use vec2 instead of scalars, to enable using dot2 instructions

* use fma instead of dot to fix Nvidia and Apple performance issues
src/ggml-vulkan/vulkan-shaders/mul_mm.comp
src/ggml-vulkan/vulkan-shaders/mul_mm_funcs.comp
src/ggml-vulkan/vulkan-shaders/types.comp
src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp