]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
CUDA: use mma FA kernel for gqa > 4 on RTX 4000 (llama/15035)
authorJohannes Gäßler <redacted>
Sat, 2 Aug 2025 14:37:08 +0000 (16:37 +0200)
committerGeorgi Gerganov <redacted>
Mon, 18 Aug 2025 17:30:45 +0000 (20:30 +0300)
commitd6e7315717a3da9cdee9c7aee9ea0ab95cabfc53
treef0ff891e3e86d1a5efb4f427ef5506dbd0ecd1fe
parenta3123e105b45c2fd2306e794c4b27fd6c0c0fa06
CUDA: use mma FA kernel for gqa > 4 on RTX 4000 (llama/15035)
ggml/src/ggml-cuda/fattn.cu