[petsc-users] matzeroentries usage

[Matthew Knepley]

On Tue, Oct 30, 2012 at 6:36 PM, NAN ZHAO <zhaonanavril at gmail.com> wrote:

> Dear all,
>
> I need to solve Ax=b in a iteratively, and the A is b is ajusted in each
> iteration. I am using matzeroentries to zeroout the values in A (A is in
> MPIBAIJ format). Somehow, I found this cause me some unexpected wrong
> solutions. I do not want to retain the non-zero structures in A when I try
> to zero it. Does anyone have a idea to do that. Or I need to destroy the A
> matrix each time I give new values to it (the non-zero structure is
> changing).
>

If the nonzero structure of A is changing, there is really no advantage to
keeping it (unless you can see one). I would
recreate it and call KSPSetOperators() again.

   Matt


>
> Thanks,
> Nan
>



-- 
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/20121030/cdb667d5/attachment.html>