]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
CUDA: fix overflow in MMA kernel without stream-k (llama/17939)
authorJohannes Gäßler <redacted>
Fri, 12 Dec 2025 16:43:58 +0000 (17:43 +0100)
committerGeorgi Gerganov <redacted>
Thu, 18 Dec 2025 06:20:56 +0000 (08:20 +0200)
commitfeb856d4a1f8cc30772f97db5583f2006b7d374c
tree6f1a5f6277e3f51cd3509396f7c4cde77ef559b8
parentdb1fcd958fb4ab2c83328500e0796dcf1ea178b4
CUDA: fix overflow in MMA kernel without stream-k (llama/17939)
ggml/src/ggml-cuda/fattn-common.cuh
ggml/src/ggml-cuda/fattn-mma-f16.cuh