Daniel,

I made the changes you suggested and manually imported the solvers from
the fipy.solvers library. I checked the results for both the scipy and
pysparse versions of the CGS, GMRES, LU, and PCG solvers. All of these
solvers produced the same correct solution when called in serial and
produced the same incorrect solution when called in parallel. This seems
to imply that the issue may not be with the solver itself. Though I am not
sure what else could be affected by the parallel environment.

As for testing the Trillinos solvers, I do not have Trillinos installed.
This is because I was originally planning on making parallel calls to the
usual solvers rather than using the parallelized Trillinos solvers.

Thanks,
Tony

_______________________________________________
fipy mailing list
[email protected]
http://www.ctcms.nist.gov/fipy
  [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]

Reply via email to