[AMD][Quantization] Add TritonScaledMMLinearKernel since int8 is broken for AMD#12282

mgoin merged 4 commits intovllm-project:mainvllm-project/vllm:mainfrom

rasmith:ransmith_triton_scaledmm_linear_kernelrasmith/vllm:ransmith_triton_scaledmm_linear_kernelCopy head branch name to clipboard