]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 5587e57)
vulkan : fix Qantized Mat-Vec Mul on AMD GPUs for ncols < 64 (#8855)
author0cc4m <redacted>
Mon, 5 Aug 2024 05:52:55 +0000 (07:52 +0200)
committerGitHub <redacted>
Mon, 5 Aug 2024 05:52:55 +0000 (08:52 +0300)
commit064cdc265fb63590c7c8f04a609d36ef200d55a7
treee9e62c0e72eb2a2157c6745b73e3507c375e62e6tree
parent5587e57a76630651752031223cc7024cb32cf308commit | diff
vulkan : fix Qantized Mat-Vec Mul on AMD GPUs for ncols < 64 (#8855)

* Fix Vulkan mul mat vec invalid results when ncols < warp size

* Only run backend ops mul mat vec block size test if block size not already covered
ggml/src/vulkan-shaders/mul_mat_vec.comp diff | blob | history
tests/test-backend-ops.cpp diff | blob | history
Packaging of ggml-org/llama.cpp