ggml-cuda : optimize mmvq nwarps for Pascal DP4A - #25479

ggml-cuda : optimize mmvq nwarps for Pascal DP4A#25479

animeshsri14 wants to merge 1 commit into

ggml-org:masterggml-org/llama.cpp:masterfrom

animeshsri14:perf/pascal-dp4a-mmvq-nwarpsanimeshsri14/llama.cpp:perf/pascal-dp4a-mmvq-nwarpsCopy head branch name to clipboard