Implement _mm512_permutex2var_epi64 shim

[elichai]

This implements the _mm512_permutex2var_epi64 intrinsic, which is vpermi2var.q.512, which permutes 64-bit elements from two input vectors according to an index vector, producing a destination vector.

[rustbot]

Thank you for contributing to Miri! A reviewer will take a look at your PR, typically within a week or two.
Please remember to not force-push to the PR branch except when you need to rebase due to a conflict or when the reviewer asks you for it.

[elichai]

Improved the test code, and I've ran it on a real x86-64 machine(AMD EPYC 9B45) with AVX512:
rustc ./tests/pass/shims/x86/intrinsics-x86-avx512.rs -C target-feature=+avx512f,+avx512vl,+avx512bw,+avx512bitalg,+avx512vpopcntdq,+avx512vnni -o avx512-test && ./avx512-test

[elichai]

I also have an implementation of _mm512_permutex2var_epi8 built on top of this permute2 function, should I add it here in a separate commit or open a new PR after this merges?
Thanks!

[RalfJung]

Please slow down a little, my review queue is already overloaded. :)

[elichai]

Please slow down a little, my review queue is already overloaded. :)

Hehe yeah I'm sorry, don't worry, I only needed these 3 intrinsics to run my code with miri :)
It's very cool that miri supports so many shims <3

[rustbot]

This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[RalfJung]

And the remaining bits? Are they ignored?

View changes since the review

[elichai]

AFAIU from https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_permutex2var_epi64&expand=4250&ig_expand=5002 it seems like bits [63:4] are ignored

[RalfJung]

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.