[petsc-users] Using SuperLU_Dist with petsc-3.0.0-p12: Necessary to use -pc_type lu ??

[Matthew Knepley]

On Tue, Jan 25, 2011 at 10:13 PM, Gaurish Telang <gaurish108 at gmail.com>wrote:

> Hi I have configured petsc-3.0.0-p12 along with SuperLU_Dist.
>
> When we use SuperLU as an external solver the manual saya that we must pass
> -pc_factor_mat_solver_package superlu_dist as an option.
>
> I notice that when we provide -pc_type lu the example runs smoothly. but
> without this option, -log_summary says
>
> WARNING! There are options you set that were not used!
> WARNING! could be spelling mistake, etc!
> Option left: name:-pc_factor_mat_solver_package value: superlu_dist
>
> Does this mean -pc_type lu must always be provided as an option when using
> SuperLU_Dist????
>

You must tell the PC to do LU factorization in order for the solver package
to be meaningful. One
way to do that is -pc_type lu.


> Also there was a bug when using SuperLU_Dist along with -ksp_view in
> petsc-3.1p5. I dont have the same problem when using petsc-3.0.0-p12 . Are
> there any reported problems when using petsc-3.0.0p12
> with SuperLU_Dist


None that I know of.

  Matt


>
> Gaurish
>



-- 
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/20110125/be2fdb43/attachment.htm>