Allow multiple OpenBLAS OpenMP calls to run in parallel

[manopapad]

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• 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.

This PR adds the required build flag to allow OpenMP-capable builds of OpenBLAS to run multiple OpenMP calls at the same time, see:

https://github.com/xianyi/OpenBLAS/blob/28a24a4d4fe3d9cd838c37504dee2493bc10a5e5/Makefile.rule#L93-L98

We would like this functionality in https://github.com/nv-legate/cunumeric, to allow us to instantiate a separate OpenMP group per NUMA domain.

NUM_PARALLEL=32 is the setting we are currently using in cuNumeric (we are building OpenBLAS from source, but would much prefer to be using the conda package), so I just copied that. If you think that is too high, we would be happy with 8, or even 2 if need be. AFAICT the main effect of increasing this number is increased internal memory usage (unclear how much):
https://github.com/xianyi/OpenBLAS/blob/28a24a4d4fe3d9cd838c37504dee2493bc10a5e5/driver/others/blas_server_omp.c#L72-L77
but no change in behavior when no more than 1 concurrent OpenMP calls are active at any time.

I don't know much about the build number versioning above, please let me know how this should be updated.

[conda-forge-linter]

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) and found it was in an excellent condition.

[isuruf]

@conda-forge-admin, rerender

[jjerphan]

Dear @manopapad,

This seems to be a really useful contribution. What is the status of this PR exactly?