Update defaults for distributed multigrid

[pratikvn]

This PR updates the defaults for distributed multigrid for the smoother and the coarse solver. It also adds a test for the distributed pgm preconditioned solver

[MarcelKoch]

@pratikvn I will fix the CI issues to push this for the release.

[MarcelKoch]

One thing to note is that the added tests are EXTREMELY slow. On nla-gpu the preconditioned test are 100x slower than the unpreconditioned tests. But I don't think this PR should fix performance issues, so I will just leave it as it is.

[upsj]

slow as in long runtime or as in slow convergence (or both)? In any case, I think we should reconsider whether to merge this PR in this state then

[MarcelKoch]

slow as in long runtime or as in slow convergence (or both)? In any case, I think we should reconsider whether to merge this PR in this state then

definitely runtime, but I think the number of iterations is also quite bad.

IMO this PR should still continue, since otherwise the user will get segfaults and the like. Even sequentially our mg is not very good, so it wouldn't be much of a difference to that.

[yhmtsai]

The slow convergence is expected in the 3pt stencil because the PGM does not handle the equal-weight matrix well currently. The corasening matrix might almost like the original matrix (n -> n - 10 for example)

[upsj]

Alright, then I withdraw my objection ;) that's an issue we should definitely tackle soon though. I think with some deterministic or randomized tie-breaking we should be able to handle that, create at least a 25%-33% reduction per level (haven't thought the worst case through entirely)

[pratikvn]

Given that we are testing that the solver + preconditioner work, I think we should just set the factory parameter with_deterministic(false) for the PGM factory. That should speedup the convergence.

[MarcelKoch]

Given that we are testing that the solver + preconditioner work, I think we should just set the factory parameter with_deterministic(false) for the PGM factory. That should speedup the convergence.

Can't confirm that this has any effect (on nla-gpu).

[sonarcloud[bot]]

Quality Gate failed

Failed conditions
B Maintainability Rating on New Code (required ≥ A)

See analysis details on SonarCloud

Catch issues before they fail your Quality Gate with our IDE extension SonarLint

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 65.11628% with 15 lines in your changes are missing coverage. Please review.

Project coverage is 89.97%. Comparing base (5b18510) to head (1fc3d7d). Report is 52 commits behind head on develop.

Files	Patch %	Lines
test/matrix/csr_kernels2.cpp	0.00%	11 Missing :warning:
core/solver/multigrid.cpp	50.00%	3 Missing :warning:
test/mpi/solver/solver.cpp	95.83%	1 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## develop #1612 +/- ## =========================================== - Coverage 90.01% 89.97% -0.04% =========================================== Files 736 752 +16 Lines 59831 60445 +614 =========================================== + Hits 53857 54387 +530 - Misses 5974 6058 +84 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.