[petsc-users] help needed for -snes_mf_operator

[Jed Brown]

On Wed, Apr 25, 2012 at 15:03, Dominik Szczerba <dominik at itis.ethz.ch>wrote:

> > This is not what I said. Look closely at the above. I said the
> > preconditioner you give
> > back when using -snes_mf_operator is not right. This has nothing to do
> with
> > normal
> > operation.
>
> I am sorry, I do not follow what this might possibly mean for me.


The operator you provided for preconditioning is not close to the finite
differenced operator. I.e. your residual evaluation and Jacobian assembly
are different to first order, probably one or both is buggy.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20120425/4fc6bf4d/attachment.htm>