Hi deal.ii community, I am having scalability issues with solvers for H(curl)-H(div) problems and H(grad)-H(curl) problems in 3D with essential boundary conditions (in a constraint matrix). Even for moderately sized problems (200K dofs) the (SparseILU preconditioned Schur-complement CG) solver needs too many iterations - depending on the coefficients between 500 and 10000. Multigrid in these spaces is (as far as I know problematic).
So I was wondering if there is a PetSc interface to MUMPS (parallel sparse direct solver) or if anyone has an idea for more efficient solvers for such problems. I am not an expert in linear solvers but their efficiency is (as so often) crucial here. I would be grateful for any idea... Cheers, Konrad -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to dealii+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
