Cao E

@mikaylagawarecki Could you please review this PR ? We would suggest it to be in PT 2.1 if possible. Thank you.

@cpuhrsch Could you please review this PR again ? Thank you.

@pytorchbot merge

@malfet The PR itself is ready, but too many UTs failed with `ATEN_CPU_CAPABILITY=default` or `simdlen=0` since they were not covered before. https://github.com/pytorch/pytorch/issues/126372, https://github.com/pytorch/pytorch/issues/126262. Take `test/inductor/test_cpu_select_algorithm.py` as instance: ``` ..... ======================================================================...

@malfet This PR is ready to be reviewed.

@peterbell10 Could you please review this PR ? Thanks.

@pytorchbot merge

@pytorchbot merge

This PR depends on https://github.com/pytorch/pytorch/issues/124245