<div>Hi,</div>
<div> </div>
<div>I tried to compile PETSc again and using --download-hypre=1. My command given is </div>
<div> </div>
<div>./config/configure.py --with-vendor-compilers=intel --with-blas-lapack-dir=/lsftmp/g0306332/inter/mkl/ --wit<br>h-x=0 --with-shared --with-mpi-dir=/opt/mpich/myrinet/intel/ --with-debugging=0 --download-hypre=1</div>
<div> </div>
<div>I tried twice and the same error msg appears:</div>
<div> </div>
<div>Downloaded hypre could not be used. Please check install in /nas/lsftmp/g0306332/petsc-2.3.2-p8/externalpackages/hypre-1.11.1b/linux-hypre. I've attached the configure.log for your reference.</div>
<div> </div>
<div>Thank you.</div>
<div> </div>
<div><br><br> </div>
<div><span class="gmail_quote">On 2/12/07, <b class="gmail_sendername">Barry Smith</b> <<a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a>> wrote:</span>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid"><br><br>On Mon, 12 Feb 2007, Ben Tay wrote:<br><br>> Hi,<br>><br>> I have some questions regarding the use of hypre/boomeramg:
<br>><br>> 1. Is there anything I need to change in the assembly of matrix etc besides<br>> adding -pc_type hypre -pc_hypre_type boomeramg ?<br><br> No<br>><br>> 2. Can it work in a sequential code?<br><br>
yes<br>><br>> 3. I have 2 eqns to solve - momentum and poisson. if I used the options,<br>> will both equations be solved using hypre?<br><br>yes<br><br>> Can I select which solver to<br>> solve with which equation?
<br><br>yes. For each KSP call KSPSetOptionsPrefix() for example<br>KSPSetOptionsPrefix(kspmo,"momentum); KSPSetOptionsPrefix(ksppo,"poisson");<br>then from the command line use -momentum_ksp_type gmres -poisson_ksp_type cg
<br>-momentum_pc_type lusomething etc. For any solver option.<br><br> Barry<br><br>><br>> Thank you.<br>><br>><br>> On 2/12/07, Barry Smith <<a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a>> wrote:
<br>> ><br>> ><br>> > hypre/boomeramg may be the way to go, especially for the Poisson<br>> > problem. -pc_type hypre -pc_hypre_type boomeramg (-help for lots of<br>> > tuning options.).<br>> >
<br>> > Barry<br>> ><br>> ><br>> > On Sun, 11 Feb 2007, Ben Tay wrote:<br>> ><br>> > > Well,<br>> > ><br>> > > I am simulating unsteady flow past a moving airfoil at Re~10^4. I'm
<br>> > using<br>> > > fractional step FVM, which means that I need to solve a momentum and<br>> > poisson<br>> > > equation.<br>> > ><br>> > > To reach a periodic state takes quite a few hours and so I'm trying to
<br>> > find<br>> > > ways to speed up the process. I thought parallelizing the code would<br>> > help<br>> > > but it seems like it's not the case.<br>> > ><br>> > > I'm now trying out different types of solver/preconditioner available on
<br>> > > PETSc to assess their performance. Is there other external solvers,<br>> > which<br>> > > PETSc interfaces, which are recommended? I'm thinking of using multigrid<br>> > to<br>> > > solve the poisson eqn... wonder if hypre/BoomerAMG etc would help...
<br>> > ><br>> > ><br>> > > On 2/11/07, Lisandro Dalcin <<a href="mailto:dalcinl@gmail.com">dalcinl@gmail.com</a>> wrote:<br>> > > ><br>> > > > On 2/10/07, Ben Tay <
<a href="mailto:zonexo@gmail.com">zonexo@gmail.com</a>> wrote:<br>> > > > > In other words, for my CFD code, it is not possible to parallelize<br>> > it<br>> > > > > effectively because the problem is too small?
<br>> > > > ><br>> > > > > Is these true for all parallel solver, or just PETSc? I was hoping<br>> > to<br>> > > > reduce<br>> > > > > the runtime since mine is an unsteady problem which requires many
<br>> > steps<br>> > > > to<br>> > > > > reach a periodic state and it takes many hours to reach it.<br>> > > ><br>> > > > Can you describe your specific application and how are you solving it?
<br>> > > > As Barry said, your need-for-speed is not likely to be solved by<br>> > > > running in parallel.<br>> > > ><br>> > > ><br>> > > > --<br>> > > > Lisandro Dalcín
<br>> > > > ---------------<br>> > > > Centro Internacional de Métodos Computacionales en Ingeniería (CIMEC)<br>> > > > Instituto de Desarrollo Tecnológico para la Industria Química (INTEC)
<br>> > > > Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET)<br>> > > > PTLC - Güemes 3450, (3000) Santa Fe, Argentina<br>> > > > Tel/Fax: +54-(0)342-451.1594<br>> > > >
<br>> > > ><br>> > ><br>> </blockquote></div><br>