[petsc-dev] tests/benchmarks

[Matthew Knepley]

On Tue, Feb 21, 2012 at 10:31 AM, Gerard Gorman <g.gorman at imperial.ac.uk>wrote:

> Gerard Gorman emailed the following on 21/02/12 16:22:
> > Hi
> >
> > I would like to harvest a selection of typical use cases for
> > benchmarking PETSc/OpenMP. Ideally they would have features such as
> > preallocating matrices rather than adding one at a time, be easy to
> > configure for different problem sizes etc. Has anyone already put
> > together such a list - what would be considered good/best practice here?
> >
> > Cheers
> > Gerard
> >
> Further to this - do I recall someone saying they had unpushed bug fixes
> to the test cases? For example, I've tried:
>
> ggorman at cynic:~/projects/petsc/petsc-dev/src/ksp/ksp/examples/tests$
> ./ex10 -m 4
> m = 2, N=375
> [0]PETSC ERROR: --------------------- Error Message
> ------------------------------------
> [0]PETSC ERROR: Argument out of range!
> [0]PETSC ERROR: New nonzero at (6,119) caused a malloc!
> [0]PETSC ERROR:
> ------------------------------------------------------------------------
> [0]PETSC ERROR: Petsc Development HG revision:
> c7afeaa80089c71502abd0b0aca07e7ea2f3d431  HG Date: Tue Feb 21 09:48:53
> 2012 -0600
> [0]PETSC ERROR: See docs/changes/index.html for recent updates.
> [0]PETSC ERROR: See docs/faq.html for hints about trouble shooting.
> [0]PETSC ERROR: See docs/index.html for manual pages.
> [0]PETSC ERROR:
> ------------------------------------------------------------------------
> [0]PETSC ERROR: ./ex10 on a arch-linu named cynic by ggorman Tue Feb 21
> 16:28:33 2012
> [0]PETSC ERROR: Libraries linked from
> /home/ggorman/projects/petsc/petsc-dev/arch-linux2-c-debug/lib
> [0]PETSC ERROR: Configure run at Tue Feb 21 16:23:30 2012
> [0]PETSC ERROR: Configure options
> [0]PETSC ERROR:
> ------------------------------------------------------------------------
> [0]PETSC ERROR: MatSetValues_SeqAIJ() line 331 in
> /home/ggorman/projects/petsc/petsc-dev/src/mat/impls/aij/seq/aij.c
> [0]PETSC ERROR: MatSetValues() line 1142 in
> /home/ggorman/projects/petsc/petsc-dev/src/mat/interface/matrix.c
> [0]PETSC ERROR: AddElement() line 198 in src/ksp/ksp/examples/tests/ex10.c
> [0]PETSC ERROR: GetElasticityMatrix() line 130 in
> src/ksp/ksp/examples/tests/ex10.c
> [0]PETSC ERROR: main() line 41 in src/ksp/ksp/examples/tests/ex10.c
> --------------------------------------------------------------------------
> MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
> with errorcode 63.
>
> NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
> You may or may not see output from other processes, depending on
> exactly when Open MPI kills them.
> --------------------------------------------------------------------------
>
> Pretty much the only value for the problem size (ie the -m option) is
> the default.
>

Yes, the preallocation is not done right. We obviously do not use this one.

   Matt


> Cheers
> Gerard
>
>


-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20120221/1b22f4f0/attachment.html>