Address XDMF/Visit Issues

[pdmullen]

Visualizing HDF5 + XDMF files generated through develop with VisIt produces strange artifacts wherein 2D datasets often show a protruding X3 feature (which looks like a skyscraper). This MR attempts to fix such artifacting by adjusting the data dumped in HyperSlab DataItems.

Parthenon particles cannot be visualized in VisIt currently, and their corresponding mesh is not even dumped in our current XDMF files. For the life of me, I cannot get VisIt to interpret VXVYVZ type geometries for Polyvertex meshes. I find that in order to get Visit to read our Polyvertex mesh, I must recast the geometry as an array of type {{x1, y1, z2}, {x2, y2, z2}...}. Even after that change, I cannot get Visit to see the Polyvertex mesh unless it is its own XDMF file. Visit can still plot particles on top of field variables if both XDMF files are read and a correlation is instantiated (VisIt prompts users automatically for this). This MR enables users to optionally dump out swarm XDMF files, which, if enabled, also dumps restructured particle geometry in the corresponding HDF5 file. Restart files should remain unchanged.

This is gross. Feel free to shoot down this MR.

PR Summary

PR Checklist

• Code passes cpplint
• New features are documented.
• Adds a test for any bugs fixed. Adds tests for new features.
• Code is formatted
• Changes are summarized in CHANGELOG.md
• Change is breaking (API, behavior, ...)
  • Change is additionally added to CHANGELOG.md in the breaking section
  • PR is marked as breaking
  • Short summary API changes at the top of the PR (plus optionally with an automated update/fix script)
• CI has been triggered on Darwin for performance regression tests.
• Docs build
• (@lanl.gov employees) Update copyright on changed files

[Yurlungur]

Hey thanks for figuring this out! And I apologize for your suffering. I don't think this MR is the optimal implementation of this machinery. But I'm not sure I care about an optimal implementation... XDMF is forcing a bizarre structure on us and I think I'm perfectly happy to just hack in a solution rather than overengineer jumping through XDMF's hoops.

This MR attempts to fix such artifacting by adjusting the data dumped in HyperSlab DataItems.

Am I missing something? I can't find this change in the diff.

[Yurlungur]

I'm approving now to indicate my willingness to just hack this in.

[pdmullen]

Am I missing something? I can't find this change in the diff.

@Yurlungur https://github.com/parthenon-hpc-lab/parthenon/pull/1079/files#diff-31b92df33d462322a957c4750de1874912d6947009746063e9684dae98314675R515

[Yurlungur]

Thanks for pointing the change out. We should test this in Assail before merging...

[lroberts36]

I am happy to approve this if it is working for you guys, but I don't think I can provide much useful feedback since I rarely use Visit and don't know what to expect for the XDMF.