[petsc-users] Multigrid preconditioner parameters

[Feng Xing]

Hello,

Thank you for replying so quick.

It’s a variation of 3d Poisson equation ( -Delta u=f, u is unknown, f is a given function ). Mathematically, they have similar properties.  So, I didn’t precise the details to make it look like too complicated. Numerically, the discretisation is by the standard finite element method or the finite volume method.

Regards,

Feng Xing
Postdoc Inria, France


> On 29 May 2015, at 21:17, Jed Brown <jed at jedbrown.org> wrote:
> 
> Feng Xing <snakexf at gmail.com> writes:
>> I want to use the the multigrid preconditioner in Petsc/hypre library
>> to a finite element elliptic problem. But there are a lot of
>> parameters. I would to ask if some of them are important for the
>> convergence and the computation time.
> 
> Yes, some of them are important.  We can't comment without further
> information on your problem, but custom tuning of solvers for each
> problem encountered by each user is not scalable for the PETSc
> developers, so we recommend that you consult the Hypre manual (if you're
> using Hypre) and experiment with the usual tuning parameters (e.g., the
> "strong threshold" discussed in the manual).  When you have more
> experience with the problem you're solving and the behavior of the
> parameters, feel free to write with more specific questions.