]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 5c16b9c)
vulkan: mul_mat_id coopmat2 optimizations (#15546)
authorJeff Bolz <redacted>
Sun, 31 Aug 2025 07:06:43 +0000 (02:06 -0500)
committerGitHub <redacted>
Sun, 31 Aug 2025 07:06:43 +0000 (09:06 +0200)
commitc37052ab4d6d1ae73c0e90bc6e560cc6409e1311
tree8f5d438412fd2daea6c4898d7f793d6402e95797tree
parent5c16b9c87d840e4d5d55fa83c732c6b693346f40commit | diff
vulkan: mul_mat_id coopmat2 optimizations (#15546)

* vulkan: mul_mat_id coopmat2 optimizations

Add a path for when the tile fits in BN/2, similar to what we have for mul_mat.

Only call fetch_scales/store_scales once per QUANT_K block, and once at the
beginning in case start_k is not aligned.

* Also add a path for BN/4 - worth a couple more percent
ggml/src/ggml-vulkan/ggml-vulkan.cpp diff | blob | history
ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp diff | blob | history
Packaging of ggml-org/llama.cpp