Set -DLIBOMP_FORTRAN_MODULES=ON on windows

[h-vetinari]

If we want to build OpenBLAS with OpenMP support with flang (c.f. conda-forge/openblas-feedstock#115), we'll need to package the Fortran bits of OpenMP, otherwise we run into:

[...]
[3646/19184] Building Fortran object CMakeFiles\LAPACK_OVERRIDES.dir\lapack-netlib\SRC\ssytrd_2stage.f.obj
[3647/19184] Building Fortran object CMakeFiles\LAPACK_OVERRIDES.dir\lapack-netlib\SRC\ssytrd_sb2st.F.obj
FAILED: CMakeFiles/LAPACK_OVERRIDES.dir/lapack-netlib/SRC/ssytrd_sb2st.F.obj 
%BUILD_PREFIX%\Library\bin\flang.exe -I%SRC_DIR%\lapack-netlib\SRC -I%SRC_DIR%\lapack-netlib\LAPACKE\include -fopenmp -fopenmp -ffixed-line-length-72 -o CMakeFiles\LAPACK_OVERRIDES.dir\lapack-netlib\SRC\ssytrd_sb2st.F.obj -c CMakeFiles\LAPACK_OVERRIDES.dir\lapack-netlib\SRC\ssytrd_sb2st.F-pp.f
error: Semantic errors in CMakeFiles\LAPACK_OVERRIDES.dir\lapack-netlib\SRC\ssytrd_sb2st.F-pp.f
D:\\bld\\openblas_1719536014536\\work\\lapack-netlib\\SRC\\ssytrd_sb2st.F:237:11: error: Cannot read module file for module 'omp_lib': Source file 'omp_lib.mod' was not found
        use omp_lib
            ^^^^^^^

Flang 18 is going to be unusable to build OpenBLAS, but I'd like to open this for discussion/review already, so we can figure this out before LLVM 19 releases.

PS. This is rebased from the dev branch, with which I was able to get CI green in conda-forge/openblas-feedstock#115.

[conda-forge-webservices]

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.

[h-vetinari]

This is quite curious. We cannot build the fortran bits of OpenMP 18 with flang 18, but it works if either of the versions is bumped to 19. In this PR we bump the flang version, on the dev branch it's the other way 'round.

In any case, even if we don't merge this for OpenMP 18, I want to enable this for OpenMP 19, so I'd appreciate feedback @conda-forge/openmp

[h-vetinari]

Gentle ping on this @conda-forge/openmp

[isuruf]

Can we move the .mod file to a different output? The cycles are also there in linux if we switch _openmp_mutex.

[isuruf]

The .mod files are specific to the gfortran version (maybe not the major version, but the soversion. not sure), so it's not possible to remove these.

[conda-forge-admin]

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

I wanted to let you know that I linted all conda-recipes in your PR (recipe/meta.yaml) and found some lint.

Here's what I've got...

For recipe/meta.yaml:

• The build section contained an unexpected subsection name. build is not a valid subsection name.
• The build section contained an unexpected subsection name. host is not a valid subsection name.

[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.

[h-vetinari]

Can we move the .mod file to a different output?

@isuruf, I've rebased this PR on #144 and added the output you wanted - I think I've incorporated all the feedback so far on this here. I still want to clean up the commit history better, but the content of the changes should be OK now. As such, I'd ask you to please review this at your earliest convenience.

[h-vetinari]

Gentle ping @isuruf

[h-vetinari]

Thanks for the review; should be fixed now.

[h-vetinari]

Good now?

[h-vetinari]

I think we're close here @isuruf, can you please help me wrap this up?