Re: Bumping CPU baseline to x86-64-v2
Ralf Gommers via NumPy-Discussion <[email protected]>
| Newsgroups | gmane.comp.python.numeric.general |
|---|---|
| Message-ID | <CABL7CQjgrwG8iy7BWjiXqAS83oSABMgWXH+t7qPbV-HcFL87Gw@mail.gmail.com> |
On Sun, May 18, 2025 at 9:40 AM Jerome Kieffer <[email protected]> wrote: > Hi, > > First I would like to highlight that "X86_V2" is very specific to the > implementation in numpy, basically this implies the support of all 128 > bits SIMD extensions, i.e. SSE1,2,3&4, but does apparenty not requires > any 256bit SIMD (AVX). > It isn't specific to NumPy. These are the psABI levels ( https://gitlab.com/x86-psABIs/x86-64-ABI), which is the closest thing there is to a standard for SIMD instruction sets. Here are some examples of Linux distros moving to x86-64-v2: - Fedora 42: https://fedoraproject.org/wiki/Changes/Optimized_Binaries_for_the_AMD64_Architecture_v2 - RHEL 9: https://developers.redhat.com/blog/2021/01/05/building-red-hat-enterprise-linux-9-for-the-x86-64-v2-microarchitecture-level# Cheers, Ralf _______________________________________________ NumPy-Discussion mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ Member address: [email protected]