[matmul kernel] [nvfp4] Use flex ctx out scale - to support tensor scale with nvfp4 output#9854

tristan-oai merged 6 commits into

triton-lang:maintriton-lang/triton:mainfrom

tristan-oai:tristan/mx-tensor-scaletristan-oai/triton:tristan/mx-tensor-scaleCopy head branch name to clipboard