Rearranging calculations in quantized linear tiled implementaiton to reduce register usage and gain performance.#15111

meta-codesync[bot] merged 1 commit intopytorch:mainpytorch/executorch:mainfrom

trivedivivek:export-D84608770trivedivivek/executorch:export-D84608770Copy head branch name to clipboard