no fast math

[Yurlungur]

PR Summary

I just discovered that in the guts of our cmake we're setting essentially -Ofast. I don't think we want to be doing that, even if it works in the code. This disables it.

PR Checklist

• Adds a test for any bugs fixed. Adds tests for new features.
• Format your changes by using the make format command after configuring with cmake.
• Document any new features, update documentation for changes made.
• Make sure the copyright notice on any files you modified is up to date.
• After creating a pull request, note it in the CHANGELOG.md file.
• LANL employees: make sure tests pass both on the github CI and on the Darwin CI
• If ML was used, make sure to add a disclaimer at the top of a file indicating ML was used to assist in generating the file.
• If Agentic AI was used, have the AI generate a "proposed changes" markdown file and store it in the plan_histories folder, with a filename the same as the MR number.

If preparing for a new release, in addition please check the following:

• Update the version in cmake.
• Move the changes in the CHANGELOG.md file under a new header for the new release, and reset the categories.
• Maintainers: ensure spackages are up to date:
  • LANL-internal team, update XCAP spackages
  • Current maintainer of upstream spackages, submit MR to spack

[dholladay00]

I thought fast math just used cuda intrinsic special functions but did not affect assiociativity and other things that -Ofast does.

[Yurlungur]

I thought fast math just used cuda intrinsic special functions but did not affect assiociativity and other things that -Ofast does.

Oh then perhaps I misunderstood what this flag did and we should turn it back on. At least in singularity-eos. In singularity-opac it was on for non-cuda builds which obviously wasn't desirable.