Support SIMD Histogram Subroutines on aarch64

[wx257osn2]

Currently SIMD histogram subroutines are written with simdlib and AVX2 intrinsics .
This PR adds some functions to simdlib and removes AVX2 intrinsics from SIMD histogram subroutines, so faiss with this PR can execute histogram using ARM SIMD on aarch64.

[wx257osn2]

Ah, ok, this PR depends on #2448. I'll re-send this PR after #2448 will be merged. Sorry for sending you an undesirable notification... 🙇

[mdouze]

The code looks good.
However, the histogram functionality is not used in the current version of the FastScan implementation, therefore I am not sure there is test coverage for the function. Could you check that? And otherwise add a test?

[mdouze]

Ah right, it is tested in
https://github.com/facebookresearch/faiss/blob/151e3d7be54aec844b6328dc3e7dd0b83fcfa5bc/tests/test_partition.py
so that looks fine.

[facebook-github-bot]

@mdouze has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[mdouze]

I have trouble landing this, I may need to make a new PR

[wx257osn2]

@mdouze Is there any action that I should do, like closing this and making a new PR with a new branch, or something like that?

[mdouze]

If you could rebase this PR on the latest Faiss it would be helpful. Thansk.

[wx257osn2]

@mdouze Sorry, I thought it was rebased by the "Update branch" button until just now (the button had been MERGING main branch into this branch, but not rebasing this branch into main) 🙇 I have rebased it, so would you import it again? thanks

[facebook-github-bot]

@mdouze has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[mdouze]

Thanks, second attempt...

[facebook-github-bot]

@mdouze merged this pull request in 5dcd282.