MNT: Support up to microarch_level 2

[matthewfeickert]

• Microarch level 2 should be generally useful, but it is unclear if level 3 actually is or not, so for the time being drop down to supporting up to level 2.
• Bump build number.

c.f. scientific-python/faster-scientific-python-ideas#11 (comment) for another example of level 3 being unclear at the moment.

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• [N/A] Reset the build number to 0 (if the version changed)
• Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• Ensured the license file is being packaged.

[matthewfeickert]

@conda-forge-admin, please rerender

[conda-forge-admin]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe/meta.yaml) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe/meta.yaml:

• ℹ️ The recipe is not parsable by parser conda-souschef (grayskull). This parser is not currently used by conda-forge, but may be in the future. We are collecting information to see which recipes are compatible with grayskull.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/12422478559. Examine the logs at this URL for more detail.}

[matthewfeickert]

cc @lgray @ikrommyd given the discussion on Slack.

[matthewfeickert]

@lgray @jpivarski from conda-forge/microarch-level-feedstock#12 (comment) people seem to indicate they don't view level 2 as being sufficiently useful. Unless there's a clear reason to provide level 2 I guess we should instead look at doing a study of level 1 vs. level 3.

[ikrommyd]

I don't personally have any strong opinions. Now that I know how to control which one to install, I'm happy. If there is a speed with v3 or v4, I wouldn't mind supporting those. If there is no real benefit, I would also be happy with just supporting v1 only.

[lgray]

v3 will bring a speedup on the modern architectures, the question for us is really one of managing UX and making sure it doesn't hit people in the face randomly.

v2 seems to only have some impact on strings (and the performance changes are debatable at best): https://stackoverflow.com/questions/46762813/how-much-faster-are-sse4-2-string-instructions-than-sse2-for-memcmp

Still I'd err on the side of more options for the compiler to work with to give us better compiled code.
I thought there were more registers, etc. that came in the latest sse versions, not just instructions. Ah well.

So, in the end v1 is probably enough, if we can manage a stable UX for v3 we should use it!