AMGCL wrapper?

[learning-chip]

https://github.com/ddemidov/amgcl

Has OpenMP/MPI/CUDA implementations, similar to Hypre but much more lightweight. Already has a Python interface (see test)

[ChrisRackauckas]

Yes, that would make sense here.

[j-fu]

WIP: see JuliaPackaging/Yggdrasil#7896

AMGCL is heavily templated. The best path forward would be a complete re-implementation in Julia which is over my time budget in the moment. That would be a great project for anyone wanting to learn precondtioning, though. In that case, the iterative solver part probably could be taken from Krylov.jl .

So for now any calls from julia need to go via a c library with pre-instantiated code, this is the above mentioned PR. Still quite bit flexibility though due to the "runtime interface".
Makes scalar and point block variants available for OpenMP in the moment, but misses e.g. saddle point stuff and CUDA.

Will first register AMGCLWrap.jl, then make a PR over here.

[j-fu]

Ok, AMGCLWrap.jl is there.

It appears to be quite straightforward to produce an AMGMCLWrapLinearSolveExt via the LinearSolveFunction API.

Anything speaking for an LinearSolveAMGCLWrapExt instead ?

[ChrisRackauckas]

I don't care either way. However it's done, I think it needs a note in the LinearSolve.jl docs and with that the details of how the wrapper is done doesn't matter to the end user.

[j-fu]

Ok so will use the first version because it is easy to maintain for me. After getting some feeling how stable it works in some of my projects I will make a PR to the docs.

[briochemc]

@j-fu if you have one handy, can you provide a MWE of using LinearSolve.jl with AMGCL.jl? (Also maybe some tests can be added here?)

[j-fu]

For the time being, see the test examples in https://github.com/j-fu/AMGCLWrap.jl/blob/b9513953f5188f223079978eb3430a5a043623d0/test/runtests.jl#L225 and following lines. You can also search for LinearSolve in the AMGCLWrap docs.

Nice thing about the precs interface of LinearSolve.jl is that preconditioner implementation does not depend on LinearSolve, and we also don't need an extension.

Note to self: make a PR to LinearSolve docs. May be @mariekeosw can help a bit with this.