[petsc-users] problems using MUMPS as a linear solver

[Dominik Szczerba]

I have figured out that I need -ksp_type preonly to get rid of fgmres.
Petsc is very flexible!
Thanks
Dominik

On Fri, Nov 25, 2011 at 12:45 PM, Dominik Szczerba <dominik at itis.ethz.ch> wrote:
> I am using MUMPS as a linear solver in my custom non-linear procedure
> (basically a loop).
> I specify the following options:
>
> -pc_type lu -pc_factor_mat_solver_package mumps -mat_mumps_icntl_4 3
> -mat_mumps_icntl_28 2 -mat_mumps_icntl_29 2 -ksp_monitor_true_residual
> -ksp_converged_reason -ksp_view
>
> I have noticed, that this does not actually do what I expect: I just
> want to unconditionally solve my system of equations with MUMPS, but I
> see that it actually is only a preconditioner. Gmres is in charge,
> performing iterations. I tried -ksp_max_it 1 but then I get divergence
> because residue reduction is somehow very poor (?), which does not
> make sense, because I expect the system be solved down to the epsilon.
>
> So is it somehow possible to just solve my Ax=b with MUMPS?
>
> Regards,
> Dominik
>