Add Generic Rosenbrock Algorithm

[dennisYatunin]

PULL REQUEST

Purpose and Content

This PR adds Rosenbrock methods to ClimaTimeSteppers. Some of these methods can work with limiters (depending on the method coefficients), but that requires specifying multiply! and set_Δtγ! in addition to linsolve!. The Rosenbrock methods support updating the Jacobian once per timestep, or once every n timesteps (the latter is not possible with OrdinaryDiffEq).

Benefits and Risks

This PR will allow us to compare the performance of our Rosenbrock23 implementation to that of OrdinaryDiffEq's. The new methods have been tested with convergence unit tests on a simple linear problem (those that support the increment formulation in addition to the tendency formulation have been tested with ForwardEulerODEFunction as well). This PR also involves adding some new tests from ARKode, as requested by Tapio, but those are not properly working yet.

PR Checklist

• This PR has a corresponding issue OR is linked to an SDI.
• I have followed CliMA's codebase contribution and style guidelines OR N/A.
• I have followed CliMA's documentation policy.
• I have checked all issues and PRs and I certify that this PR does not duplicate an open PR.
• I linted my code on my local machine prior to submission OR N/A.
• Unit tests are included OR N/A.
• Code used in an integration test OR N/A.
• All tests ran successfully on my local machine OR N/A.
• All classes, modules, and function contain docstrings OR N/A.
• Documentation has been added/updated OR N/A.

[charleskawczynski]

Wait, we already had a Rosenbrock time stepper? How is what was added different? Can this please be elaborated in the description? The previous implementation (that it looks like @simonbyrne implemented?) looks much simpler

[tapios]

I have a similar question, perhaps phrased slightly differently: Why does enabling limiters necessitate this much more complexity than the straightforward Rosenbrock implementation in the older code?