[petsc-users] Petsc with ML and ILU

[Matthew Knepley]

On Fri, May 2, 2014 at 9:38 AM, Lucie Freret <
lfreret at arrow.utias.utoronto.ca> wrote:

> Hello,
>
> I would like to solve linear systems using Gmres preconditioned by ML and
> use ILU(0) on all levels (mg_coarse and mg_levels_x).
> As I have MATMPIAIJ matrix, I'm using
> -ksp_type gmres
> -pc_type ml
> -mg_levels_ksp_type preonly (-mg_coarse_ksp_type preonly)
> -mg_levels_pc_type asm (-mg_coarse_pc_type asm)
> but I get:
> "Running KSP of preonly doesn't make sense with nonzero initial guess"
> I tried different keyword to have a zero initial guess but unfortunately,
> I can't solve this problem.
> Should I use an other mg_levels_ksp solver of each level or is there a way
> to initialize guess on each level?
>

You want "richardson" instead of preonly, since you are doing defect
correction in MG.

   Matt


> Thanks,
> Lucie
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
>
>
>


-- 
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/20140502/13872f34/attachment.html>