]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
overview / pkg / ggml / sources / ggml / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 3571e69)
HIP: enable WMMA-MMQ INT kernels for RDNA 3 (llama/17576)
authorJiacheng (Jason) Chen <redacted>
Fri, 5 Dec 2025 08:17:37 +0000 (03:17 -0500)
committerGeorgi Gerganov <redacted>
Thu, 11 Dec 2025 13:32:54 +0000 (15:32 +0200)
commitb96dffa007fd33b8036a2961eb432c8f881fa00e
tree739cfafbb9842ac1e8a815f4b69ca09bf284ff2etree
parent3571e69bdded2e7df7c6caa6723aee83918d6ac8commit | diff
HIP: enable WMMA-MMQ INT kernels for RDNA 3 (llama/17576)

* enabled wmma instructions for most quantizations other than q2k

* fixed the last q2_k test case failure

* address comments: fix out of bound write for RDNA4, add comments after #endif

* clean up rebase: fix ne error in half2

* fix the EditorConfig CI
src/ggml-cuda/common.cuh diff | blob | history
src/ggml-cuda/mma.cuh diff | blob | history
src/ggml-cuda/mmq.cu diff | blob | history
src/ggml-cuda/mmq.cuh diff | blob | history
Packaging of ggml-org/ggml