]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: cc45f2a)
Adjust workaround for ROCWMMA_FATTN/GFX9 to only newer ROCm veresions (#19591)
authorMario Limonciello <redacted>
Mon, 16 Feb 2026 13:46:08 +0000 (07:46 -0600)
committerGitHub <redacted>
Mon, 16 Feb 2026 13:46:08 +0000 (14:46 +0100)
commit2ba9adc093127016a48cd0c5d6bf1420dafe17a6
tree514d239b475e330dddb3d5926b64fccec276fedetree
parentcc45f2ada695644c6697c0fb0e70a5e95563ad0fcommit | diff
Adjust workaround for ROCWMMA_FATTN/GFX9 to only newer ROCm veresions (#19591)

Avoids issues with ROCm 6.4.4.

Closes: https://github.com/ggml-org/llama.cpp/issues/19580
Fixes: 6845f7f87 ("Add a workaround for compilation with ROCWMMA_FATTN and gfx9 (#19461)")
Signed-off-by: Mario Limonciello (AMD) <redacted>
ggml/src/ggml-cuda/fattn-wmma-f16.cu diff | blob | history
Packaging of ggml-org/llama.cpp