[Breaking] Linear Solver Updates

[lroberts36]

PR Summary

This PR adds a number of solver related things:

• Adds a CG solver. I am not sure how useful this is since it seems to work poorly with AMR and multigrid preconditioning, but I think it is at least nice to have in there for testing.
• Adds simple diagonal preconditioners for CG and BiCGSTAB
• Adds the option to use a user defined task for prolongation during multigrid solves instead of the prolongation facilities provided by the Parthenon infrastructure. This more easily allows for physical boundary aware prolongation operators.
• Adds a number of different boundary aware prolongation operators (constant, bilinear, and Kwak) to the poisson_gmg example.
• Adds explicit setting of boundary fluxes and fluxes at refinement boundaries in the poisson_gmg example.
• Adds a base class for solvers to simplify switching between solvers for downstream codes
• Switches to stage based solvers that can include multiple separate variables as part of the state vector. This is a breaking change for any code that is using solvers downstream.

To facilitate these changes to the solvers and examples, two small changes are made to the rest of the infrastructure:

• Extends the neighbor information that @brryan put into SparsePacks to work with multigrid neighbors and include physical boundary information as well. This is a breaking change, since now physical boundaries are denoted with a large, negative level in the array containing neighbor information in SparsePack
• Adds a mechanism for getting a flat index from a multi-d Indexer.

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]

LGTM. This seems to be a pretty significantly cleanup along with the new features and better performance.

[bprather]

I've been using this as-is in KHARMA for ages now, it works for me. The code looks great, I like the simpler interface.