On Wed, Oct 26, 2011 at 23:40, Paul Sicsic <paul.sicsic at polytechnique.edu>wrote:
> I have attached the output of valgrind for a run with no options You can do this run in a debugger with the option -fp_trap. It will break at the first place where a denormal or NaN is computed. > and for a run with the options you have recommended earlier. The X you passed in is not a valid Mat (we could see that from the earlier call). More specifically, according to valgrind, it is uninitialized memory. The shift is still infinite coming into this function, which I don't understand. >From the singularity at 0, it looks like perhaps the initial time step size is being set to zero. Next time you get a stack trace, can you on ts->time_step? -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111026/81ebb901/attachment.html>
