I'll mess around and see what happens. However, since what I'm doing is solving a domain while changing the conductivity in successive iterations, I don't have an exact notion of whether there's a particular scenario that causes the failure or if its random.
Best Regards, Angus Hendrick On Mon, Jan 5, 2009 at 3:02 PM, Daniel Wheeler <[email protected]>wrote: > > Is there a revision of fipy for which the problem actually works? What > happens if you go back 10 or 20 revisions. Does it work? > > On Mon, Jan 5, 2009 at 4:48 PM, Angus Hendrick <[email protected]> > wrote: > > Hello, > > > > All the sudden I seem to be getting segmentation faults with the Trilinos > > GMRES Solver. Attached is a backtrace/memory dump. At the beginning you > > can see the "Solving... returned" pairs, which I stuck into my code right > > before and after I called the "solve" method on my equation. That the > > backtrace starts after "Solving..." but before returned indicates the > > problem happened during the solve method. > > > > I updated to the latest FiPy trunk version AFTER these started occuring, > but > > they are still happening. I have switched to the CGS and LU solvers and > so > > far have not encountered the problem with either of them. > > > > Best Regards, > > Angus Hendrick > > > > > > > > -- > Daniel Wheeler > >
