Build matrix of MPI variants

conda-forge / vtk-feedstock

A conda-smithy repository for vtk.

BSD 3-Clause "New" or "Revised" License

13 stars 64 forks source link

Build matrix of MPI variants #123

Closed xylar closed 4 years ago

xylar commented 4 years ago

Since VTK depends on libnetcdf and hdf5, it will pick only the nompi variant of MPI by default (because this has a higher build number for both libraries). However, users may try to run with libnetcdf versions with MPI support, which may cause problems.

I propose that we build with all 3 MPI variants but an alternative would be to explicitly only support nompi.

Checklist

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

conda-forge-linter commented 4 years ago

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.

xylar commented 4 years ago

@conda-forge-admin, please rerender

xylar commented 4 years ago

@conda-forge-admin, restart ci

xylar commented 4 years ago

@conda-forge-admin, please rerender

github-actions[bot] commented 4 years ago

Hi! This is the friendly automated conda-forge-webservice. I tried to rerender for you, but it looks like there was nothing to do.

xylar commented 4 years ago

@conda-forge-admin, restart ci

xylar commented 4 years ago

@conda-forge-admin, please rerender

github-actions[bot] commented 4 years ago

Hi! This is the friendly automated conda-forge-webservice. I tried to rerender for you, but it looks like there was nothing to do.

xylar commented 4 years ago

@conda-forge/vtk, anyone have objections to or questions about supporting various MPI builds of libnetcdf and hdf5? This shouldn't affect VTK directly but is important for environments that want to have VTK installed and use MPI for other packages (ESMF is one common example).

This does roughly triple the number of build variants produced in each PR, so that's a consideration for sure.

Another option, at least for now, would be to explicitly require that the nompi version of libnetcdf and hdf5 be used and not support the MPI builds. We currently don't specify, which could lead to building against nompi but a user trying to install mpich or openmpi builds and (I believe) finding they don't work.