Excellent! I am happy to say that it not converges to a result, which looks to be the same that I found in fenics. Thanks for all the help! Cheers, Francis ------------------ Francis Poulin Associate Professor Department of Applied Mathematics University of Waterloo email: fpoulin@uwaterloo.ca Web: https://uwaterloo.ca/poulin-research-group/ Telephone: +1 519 888 4567 x32637 ________________________________________ From: firedrake-bounces@imperial.ac.uk [firedrake-bounces@imperial.ac.uk] on behalf of Lawrence Mitchell [lawrence.mitchell@imperial.ac.uk] Sent: Tuesday, November 29, 2016 4:36 PM To: firedrake@imperial.ac.uk Subject: Re: [firedrake] no normal flow Boundary Conditions
On 29 Nov 2016, at 21:25, Francis Poulin <fpoulin@uwaterloo.ca> wrote:
Thanks Lawrence.
That did seem to help. Now I get another errror but now I get that it fails to converge after 0 iterations.
It suggests that I add some options. Should I try doing this in firedrake? I am not sure how to use these options but if you think it would help and I can figure out how I am happy to give it a try.
Oh, I should have said. The default LU in petsc does no pivoting and you have a whacking great zero block in your system. Try additionally with: "pc_factor_mat_solver_package": "mumps" The error you saw comes because petsc encountered a zero pivot in the factorisation stage. Lawrence _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake