[petsc-users] Any suggestions for improving the convergence rate or performance?

[Jed Brown]

Try using a multilevel method for the Poisson solve. If you use DM, you can
just do -pc_type mg -pc_mg_levels 4 (or whatever) to do geometric
multigrid. Otherwise, start with algebraic multigrid, -pc_type gamg, or
configure with ML or Hypre and -pc_type ml or -pc_type hypre. The iteration
count should be much lower, though each iteration will be somewhat more
expensive.


On Sat, Jun 30, 2012 at 12:16 AM, Bao Kai <paeanball at gmail.com> wrote:

> Hi, all,
>
> I am currently trying to solve a solute transport problem with PETSC
> and finite difference method with structured grids. Two equations are
> involved, one is a transport equation with advection and diffusion
> terms, another one is the flow equation with gravity, which is
> actually a poisson equation. Both the equations use the same solver.
>
> Since we eventually want to handle non-linear problems, we us SNES
> here to solve the equations. I use the BICGSTAB ksp linear solver and
> block Jacobi preconditioner.  All the tolerances are the default ones
> from PETSC.  The range of the solution values for transport equation
> is about 0-1100, and the range of the solution values of the flow
> equations should be between -5000-5000.
>
> The problem is that the solution of flow equations takes about ten
> times of the time than the transport equations. I am wondering if I
> should set different tolerances for the two equations.
>
> Attached please find the convergence processes of the two equations.
> Generally it takes about ten times of iterations to solve the flow
> equations than transport equations. I only do applications and do not
> have good knowledge in the solution of linear system, so I am here to
> find some suggestions on this issue.
>
> The size of the mesh is 500X500X500 (125 million unknowns ) and 512
> nodes (2048 processors ) on BlueGene/P machine are used.
>
> Any suggestions will be much appreciated.
>
> Thank you .
>
> Kai
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20120630/cd498916/attachment.html>