llama.cpp
Vectorize load instructions in dmmv f16 CUDA kernel
#9816

Merged

Commits

Vectorize load instructions in dmmv f16 CUDA kernel

agray3 committed 1 year ago
addressed comment

agray3 committed 1 year ago
Update ggml/src/ggml-cuda/dmmv.cu

agray3 committed 1 year ago

Loading