[petsc-dev] Enabling single precision for MUMPS and SUPERLU

Stefano Zampini stefano.zampini at gmail.com
Fri Sep 21 10:30:43 CDT 2012

2012/9/21 Hong Zhang <hzhang at mcs.anl.gov>

> Stefano:
> Thanks for sending us the patch. Is this for petsc-3.3 or petsc-dev?


Barry: you're the boss but I agree with Jed. It will be a very ugly code if
we duplicate a bunch of #ifdef inside mumps.c. Regarding superlu, I decided
to maintain them since they are fewer. You don't like PetscMUMPS_c? Well,
you only need to change the first lines :-)

I was thinking about using MUMPS_c (as already did for BLAS/LAPACK), but
it's too much close to xmumps_c, maybe it will interfere with their macros
in the next release.

When Hong has finished testing, I will split the patches and commit them.
Please let me know what macro name you'll choose (or the ugly choice of

There is also pastix which can be enabled, but the problem is that it
requires pt-scotch for partitioning, and in turn pt-scotch will become a
possible method to be used in MatPartitioning class. Maybe someone can
think about adding a new command line argument in package configure which
guaranties you will use a library (pt-scotch) only inside  pastix and not
within the MatPartitioning class? (maybe using self.useddirectly in

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20120921/07575880/attachment.html>

More information about the petsc-dev mailing list