"Sun, Hui" <[email protected]> writes:

> Thank you Jed. I see. 
>
> I have another question: Petsc uses multigrid as a preconditioner. How do I 
> specify the option so that it becomes a solver? Is it by doing:
> ./ex34 -pc_type mg -pc_mg_type full -ksp_type richardson -ksp_monitor_short 
> -pc_mg_levels 3 -mg_coarse_pc_factor_shift_type nonzero

Yes.

> MG as preconditioner or as solver, which one gives better performance and 
> why? 

FMG done right is a direct solver (one iteration).  Otherwise, Krylov
usually provides enough benefit to justify its modest cost, but it all
depends on what is slowing convergence.

Hammer or wrench, which is more useful and why?

Attachment: signature.asc
Description: PGP signature

Reply via email to