[petsc-users] selection of KSP solver

[Matthew Knepley]

On Tue, Oct 22, 2013 at 1:58 PM, Roc Wang <pengxwang at hotmail.com> wrote:

> Hi,
>
> I have a linear system with about 161^3=4,173,281 unknowns.  I used the
> default KSP solver(to my understand it is GMRES) to solve it. The relative
> tolerance is set as 1.266e-9. The number of iterations is 5373 and the
> computing time problem is about 23.65 mins with 128 processes.   With the
> same problem, but I selected a pre-conditioner such as hypre, the number of
> iterations can be reduced dramatically, however, the computing time
> increases somewhat.
>
> May someone give me some suggestions on selection of KSP solver to solve a
> large scale linear system faster?  Thanks a lot.
>

There are no blackbox preconditioners. You must find one that matches your
system. The right thing to do
is to lookup the literature on your problem and see what other people have
done. It should be easy to
replicate in PETSc.

   Matt

-- 
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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20131022/44c2fe75/attachment.html>