CUDA: route batch>=4 quantized matmul to MMQ on AMD MFMA hardware#23227

JohannesGaessler merged 3 commits into

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

jadenmach2:cdna-mmq-batch4jadenmach2/llama.cpp:cdna-mmq-batch4Copy head branch name to clipboard