TPL Update 2025-12-15

[chapman39]

this pr

• updates packages
  • mfem mfem/mfem@2328532 (eric's branch)
  • axom llnl/axom@52ef76c
  • sundials to 7.5.0
  • tribol llnl/Tribol@9f9d4d6
  • spack spack/spack@0154734
  • spack-packages spack/spack-packages@88ce9cc
  • enzyme to 0.0.234
  • cmake to (at least) 3.24, due to raja package https://github.com/spack/spack-packages/blob/develop/repos/spack_repo/builtin/packages/raja/package.py#L297
• removes radiuss-spack-configs submodule
• replaces +profiling with +adiak+caliper Adds a components variant to Axom's spack package axom#1740
• fix nvcc compiler errors on cuda docker container
• new host configs for LC machines + docker containers

about CUDA

• i fixed to build issues, but we run out of space on github runners, so it's still disabled
• however, the container works and is up to date, so i'll leave it in this PR

related PRs

• mfem: fix hypre version constraint spack/spack-packages#2978
• Make llvm path consistent between llvm and llvm-amdgpu in enzyme pkg spack/spack-packages#2365
• Update radiuss packages with latest releases spack/spack-packages#2558

next

• update smith cz and lido
• update tribol's docker containers to include cuda as well

Fixes #1525

[white238]

Heads up for you @chapman39 , Axom changed their default type for index:

https://github.com/llnl/axom/blob/623a1d9d139176737e2ce60c17c06cc2d2d40e89/RELEASE-NOTES.md?plain=1#L169

You will get a lot of warnings as errors.

[white238]

This is weird.

[chapman39]

i forget the exact error, but spack wasn't concretizing without this.

[white238]

Should this be removed then? Why have an external when we are building it?

[chapman39]

i believe non-cuda containers were finding openblas, which is why i figured i'd leave the external path just in case.

[white238]

All my comments were nitpicking and I don't want to hold up this being merged after someone else's review. Thanks for the hard slog @chapman39 !