On Wed, Oct 26, 2011 at 23:09, Paul Sicsic <paul.sicsic at 
polytechnique.edu>wrote:

>
>    - The dot variable is still Nan when trying to compute the IFunction.
>    - These options seem to still be giving a shift that is infinity when
>    computing the IJacobian. (In step 0 here)
>
> Yeah, this is the problem, but I can't reproduce it for other problems.
With these symptoms, I would expect memory corruption or a build/link error.
Can you confirm that the examples run correctly?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: 
<http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111026/24357ba5/attachment.html>

Reply via email to