Collin Wittenstein
Collin Wittenstein
any successions of how one should name the methods when implementing? Just Sandu2A, Sandau2B, ...?
I just realized that Methods A-C don't have an embedded method. Does it still make sense to implement them? And method D has an embedded method, but it took ~2000...
@ChrisRackauckas does this PR look good to you? Then I will start adding step_limiter to more algorithms.
I could figure out the source of the problem, but I think it is somewhere related to the initialization of the problem. When doing ```julia @show "hey from cache" @show...
The PR looks way bigger then it is, but this is only because of the (wrong?) changes in the Manifest.toml.
@CarloLucibello
I assume this is because ``my_implicit_euler()`` is able to reformulate $u^{n+1} = u^{n} + \Delta t f(u^{n+1})$ into $(I - \Delta t f) u^{n+1} = u^{n}$ using ``FunctionOperator`` to make...
> If you check `sol.stats`, is it refactorizing? My guess is that the "just let it be nonlinear" form is not recognizing that it can do the whole problem with...
This is now slowly beyond me expertise, but does this also explain the 10x difference when using indirect solvers like ``KrylovJL_GMRES()``? Because there you dont have any factorizations to store...
Maybe I am wrong, but isn't the W matrix only needed when solving nonlinear systems, and could it be that considerable time is spent evaluating it even when solving linear...