[petsc-users] MatMAIJ tests

Nunion m.skates82 at gmail.com
Tue Dec 7 20:52:20 CST 2010


Yes, creating a sister file is the approach I am taking.  I am interested in
alternate approaches (implementations) for the matrix multi-vector multiply
for improving performance.  I read that this can be achieved using the
MatMult function on matrices created with MatCreateMAIJ.  My problem happens
to be complex.

On Tue, Dec 7, 2010 at 3:50 PM, Jed Brown <jed at 59a2.org> wrote:

> On Tue, Dec 7, 2010 at 22:45, Nunion <m.skates82 at gmail.com> wrote:
>
>> What would be a fix to get around the issue with complex numbers for  ex100.c:
>> Tests vatious routines in MatMAIJ format<http://w3.pppl.gov/m3d/petsc-dev/src/mat/examples/tests/ex100.c.html>in the src/mat/examples/tests/ directory.
>> Currently it is not written for complex numbers.
>>
>
> The code in that test would not have to change, but the on-disk binary
> format would need to change so that real matrices could be loaded.  Or a
> sister file would need to be written that did use complex.  Why do you want
> to run this test with complex?
>
> Jed
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20101207/1d15e243/attachment.htm>


More information about the petsc-users mailing list