[petsc-users] Preconditioner and number of processors

[Matthew Knepley]

On Thu, Apr 17, 2014 at 2:48 PM, Xiao, Jianjun (IKET)
<jianjun.xiao at kit.edu>wrote:

> Dear developers,
>
> I find the words below in PETSc FAQ.
>
> "Most commonly, you are using a preconditioner which behaves differently
> based upon the number of processors, such as Block-Jacobi which is the
> PETSc default."
>
> I am testing the fidelity of my code. I don't want to be disturbed by the
> different numbers of procs. I know Block-Jacobi may be fast. But currently
> I don't care the speed. I need a preconditioner which behaves exactly the
> same with various procs. Which preconditioner should I use?
>

Always use a direct solver for fidelity tests, like SuperLU or MUMPS.

   Matt


> Thank you.
>
> Best regards
> JJ




-- 
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/20140417/634ded31/attachment.html>