On 25 Jul 2014, at 21:28, Jed Brown <[email protected]> wrote: > Sorry about not following up. I also find these results peculiar. > > Lawrence Mitchell <[email protected]> writes: >> So I'm sort of none-the-wiser. I'm a little bit at a loss as to why >> this occurs, but either switching to Richardson+SOR or Cheby/SOR with >> more that one SOR sweep appears to fix the problems, so I might just >> punt for now. > > What discretization and mesh is this running on?
Bog-standard P1 on a pretty much regularly meshed square domain (i.e. no reentrant corners or bad elements). > Is there something special about the decomposition with 2 subdomains? It doesn't look like it, the two subdomains are about the same size. > Are the Chebyshev estimates far from converging? So for the two-level problem, if I compute the extremal eigenvalues of the preconditioned operator being used as a smoother I get (approximately): 1 process: 0.019, 1.0 2 processes: 0.016, 1.4 3 processes: 0.016, 1.36 The eigenvalue estimates (from ksp_view) are: 1 process: 0.09, 1.01 2 processes: 0.09, 1.01 3 processes: 0.13, 1.47 When I bump to more levels, the estimates are only bad on two processes on the finest grid. So for example running with something like (for a 3 level problem): -pc_type mg -mg_levels_2_ksp_chebyshev_eigenvalues 0.09,1.4 gives me good convergence on two processes where the extremal eigenvalues on the finest grid are: 0.016, 1.39 Cheers, Lawrence
signature.asc
Description: Message signed with OpenPGP using GPGMail
