> This can happen if the linear system is very ill conditioned. You can run > the direct solve inside -ksp_type richardson for a more classical "iterative > refinement". Using a Krylov method is generally better. Because of the way > GMRES preconditioning works, converging after one iteration still means two > preconditioner applications (so much cheaper than factorization that you > probably don't notice it). In contrast, one iteration of FGMRES only does > one preconditioner application.
Yes, the system is badly conditioned. But am I right with my (experimental) finding that the ultimate way to go is -ksp_type preonly? Will it do what I want: just solve my system exactly once using MUMPS and not Krylov? Thanks Dominik
