]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
CUDA: fix MMQ for non-contiguous src0, add tests (llama/10021)
authorJohannes Gäßler <redacted>
Thu, 24 Oct 2024 09:09:36 +0000 (11:09 +0200)
committerGeorgi Gerganov <redacted>
Sat, 26 Oct 2024 07:31:47 +0000 (10:31 +0300)
commit71b5af5920df0d6a7236dc44fbc59c71fede0fb8
tree67ca5c55e9d67db6849b72570776376ba6b83268
parentf879e97d806b5ff15f0e67fd277fa2c81c024be8
CUDA: fix MMQ for non-contiguous src0, add tests (llama/10021)

* CUDA: fix MMQ for non-contiguous src0, add tests

* revise test code
src/ggml-cuda.cu
src/ggml-cuda/mmq.cu
src/ggml.c
tests/test-backend-ops.cpp