[petsc-users] SNES failed with reason -3

[Peter Brune]

-3 is a linear solver failure.  You can see the whole list of
convergence/divergence reasons under SNESConvergedReason in
$PETSC_DIR/include/petscsnes.h

Anyhow, what kind of output do you get when running with -ksp_monitor and
-ksp_converged_reason?

- Peter


On Wed, Apr 16, 2014 at 9:57 AM, Que Cat <quecat001 at gmail.com> wrote:

> Dear Petsc-Users,
>
> I call the nonliner solver snes in my code and it said that "SNES Failed!
> Reason -3". I testes with "-pc_type lu" and it passed. I have checked the
> hand-coded Jacobian, ant it looks fine. Could you suggest any possible
> sources of the errors? THank you.
>
> Que
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20140416/6b196421/attachment-0001.html>