]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
CUDA: batch out_prod inner loop with cublasSgemmStridedBatched (llama/22651)
authorleonardHONG <redacted>
Thu, 7 May 2026 19:59:29 +0000 (03:59 +0800)
committerGeorgi Gerganov <redacted>
Thu, 14 May 2026 18:26:48 +0000 (21:26 +0300)
commit6e91ed3b338f221c51c02439abf4c4e952ebe306
treed63c2a994f1daa97e85def1700a00df32a966406
parent5fd75cda3fec6b87f724a5784479f8ff9348a7d4
CUDA: batch out_prod inner loop with cublasSgemmStridedBatched (llama/22651)

* CUDA: batch out_prod inner loop with cublasSgemmStridedBatched

* CUDA: batch out_prod inner loop with cublasSgemmStridedBatched

* CUDA: add cublasSgemmStridedBatched mapping for HIP and MUSA backends
ggml/src/ggml-cuda/out-prod.cu
ggml/src/ggml-cuda/vendors/hip.h
ggml/src/ggml-cuda/vendors/musa.h