[petsc-dev] Seeking Advice on Petsc Preconditioners to Try

Mark F. Adams mark.adams at columbia.edu
Fri Dec 23 20:41:33 CST 2011

On Dec 23, 2011, at 8:36 PM, Dave Nystrom wrote:

> Mark F. Adams writes:
>> It sounds like you have a symmetric positive definite systems like du/dt -
>> div(alpha(x) grad)u.  The du/dt term makes the systems easier to solve.
>> I'm guessing your hard system does not have this mass term and so is
>> purely elliptic.  Multigrid is well suited for this type of problem, but
>> the vector nature requires some thought.  You could use PETSc AMG -pc_type
>> gamg but you need to tell it that you have a system of two dof/vertex.
>> You can do that with something like:
>> ierr = MatSetBlockSize( mat, 2 );      CHKERRQ(ierr);
> Best I can tell, the matrix block size cannot be set using a command line
> option.  It that correct?

I believe so.

> Thanks,
> Dave
>> For the best results from GAMG you need to give it null space information
>> but we can worry about that later.
>> Mark

More information about the petsc-dev mailing list