Rohanjames1997

@pytorchmergebot merge

@ajitmaths, @gqchen, @zhxchen17, @Mortimerp9, @jeanschmidt Could one of you review this PR? (Maintainer approval required for merge) CC:@malfet, @jgong5

@malfet, Thanks for the review! Yes, all ARMv8 CPUs do support NEON. I added a separate dispatch for NEON as the default implementation was the slow path - it didn't...

@malfet thanks for pointing it out. But since this entire change revolves around the `ASIMD_FOUND` variable, I think the code wouldn't bloat on mobile devices. The reason being - `ASIMD_FOUND`...

@malfet, in addition to the above, could you let me know how I can generate a mobile build on my Linux machine? I want to check if the size of...

@malfet , a gentle reminder for the same 😄

@malfet @nikitaved , Could one of you re-review the PR? Just wanted to ensure that this PR and #105590 have not slipped through the cracks 😄

Hi @malfet @jgong5, could the Stale label be removed for this PR as well? Thanks! Any additional review comments would be greatly appreciated!

@jgong5 can you please reopen this PR?

Gentle ping! Would appreciate if this PR could be reviewed! I'd like to contribute this optimization before the code-freeze for Pytorch2.3! Happy to address any further review comments! @malfet @nikitaved