On Mon, 12 Feb 2007, Ben Tay wrote:
> Hi, > > I have some questions regarding the use of hypre/boomeramg: > > 1. Is there anything I need to change in the assembly of matrix etc besides > adding -pc_type hypre -pc_hypre_type boomeramg ? No > > 2. Can it work in a sequential code? yes > > 3. I have 2 eqns to solve - momentum and poisson. if I used the options, > will both equations be solved using hypre? yes > Can I select which solver to > solve with which equation? yes. For each KSP call KSPSetOptionsPrefix() for example KSPSetOptionsPrefix(kspmo,"momentum); KSPSetOptionsPrefix(ksppo,"poisson"); then from the command line use -momentum_ksp_type gmres -poisson_ksp_type cg -momentum_pc_type lusomething etc. For any solver option. Barry > > Thank you. > > > On 2/12/07, Barry Smith <bsmith at mcs.anl.gov> wrote: > > > > > > hypre/boomeramg may be the way to go, especially for the Poisson > > problem. -pc_type hypre -pc_hypre_type boomeramg (-help for lots of > > tuning options.). > > > > Barry > > > > > > On Sun, 11 Feb 2007, Ben Tay wrote: > > > > > Well, > > > > > > I am simulating unsteady flow past a moving airfoil at Re~10^4. I'm > > using > > > fractional step FVM, which means that I need to solve a momentum and > > poisson > > > equation. > > > > > > To reach a periodic state takes quite a few hours and so I'm trying to > > find > > > ways to speed up the process. I thought parallelizing the code would > > help > > > but it seems like it's not the case. > > > > > > I'm now trying out different types of solver/preconditioner available on > > > PETSc to assess their performance. Is there other external solvers, > > which > > > PETSc interfaces, which are recommended? I'm thinking of using multigrid > > to > > > solve the poisson eqn... wonder if hypre/BoomerAMG etc would help... > > > > > > > > > On 2/11/07, Lisandro Dalcin <dalcinl at gmail.com> wrote: > > > > > > > > On 2/10/07, Ben Tay <zonexo at gmail.com> wrote: > > > > > In other words, for my CFD code, it is not possible to parallelize > > it > > > > > effectively because the problem is too small? > > > > > > > > > > Is these true for all parallel solver, or just PETSc? I was hoping > > to > > > > reduce > > > > > the runtime since mine is an unsteady problem which requires many > > steps > > > > to > > > > > reach a periodic state and it takes many hours to reach it. > > > > > > > > Can you describe your specific application and how are you solving it? > > > > As Barry said, your need-for-speed is not likely to be solved by > > > > running in parallel. > > > > > > > > > > > > -- > > > > Lisandro Dalc??n > > > > --------------- > > > > Centro Internacional de M??todos Computacionales en Ingenier??a (CIMEC) > > > > Instituto de Desarrollo Tecnol??gico para la Industria Qu??mica (INTEC) > > > > Consejo Nacional de Investigaciones Cient??ficas y T??cnicas (CONICET) > > > > PTLC - G??emes 3450, (3000) Santa Fe, Argentina > > > > Tel/Fax: +54-(0)342-451.1594 > > > > > > > > > > > >
