]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
CUDA: no FP16 arithmetic for vector FA kernel (llama/17558)
authorJohannes Gäßler <redacted>
Fri, 28 Nov 2025 09:29:09 +0000 (10:29 +0100)
committerGeorgi Gerganov <redacted>
Thu, 11 Dec 2025 13:32:48 +0000 (15:32 +0200)
commita8a933fb1fe8b612459a8c340aed81f567d15cfa
tree9f182f6eac2c2be6d68f3f98ad67257a4c46199a
parentcf3c9a9a103e6dc5d169899cee81a2fc5cfc2225
CUDA: no FP16 arithmetic for vector FA kernel (llama/17558)
src/ggml-cuda/common.cuh
src/ggml-cuda/fattn-common.cuh
src/ggml-cuda/fattn-vec.cuh