[petsc-users] How to choose a suitable restarted m when using GMRES?

Jed Brown jedbrown at mcs.anl.gov
Sat Nov 24 10:46:29 CST 2012

On Sat, Nov 24, 2012 at 5:38 PM, w_ang_temp <w_ang_temp at 163.com> wrote:

>     So, is there no effective way to determine a suitable m? Only by doing
> lots of tests to find
> an appropriate value in practical application? Or something need to be
> noticeable of using GMRES
> with an effective way?

Just use the default until it doesn't converge well or until you run out of
memory. You're not going to have any theory here unless you know a lot of
about your problem and want to think carefully about it, neither of which
seem to be true, and even if it was, it's unlikely to be practically
important. If you want some entertainment, read Mark Embree's 2003 SIAM
Review article "The tortoise and the hare restart GMRES".

Consider using MINRES instead of GMRES.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20121124/0d341caf/attachment.html>

More information about the petsc-users mailing list