[petsc-users] Motivation for default KSP solver and PC

[Matthew Knepley]

On Mon, Sep 21, 2020 at 8:51 AM Luciano Siqueira <luciano.siqueira at usp.br>
wrote:

> Hi *,
>
> I'm experimenting with different combinations of KSP solvers and PCs and
> I don't know why GMRES/bjacobi are the default choices for CPU and
>

GMRES is chosen because it is monotonic in the 2-norm, and also faster than
any other
Krylov method in terms of iterates if you do not restart. Block-Jacobi is
trivially parallel.
I don't think anyone loves ILU(0), but it is a black-box preconditioner
without much overhead,
so it is the default.


> GMRES/icc are the default choices for GPU. Does anyone know the reason for
> that?
>

ICC is only a default if you have a symmetric matrix. I guess you had one
on the GPU.

  Thanks,

     Matt


> Thanks,
> Luciano.
>


-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener

https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20200921/2f89d9b7/attachment.html>