[petsc-dev] Workaround in MatTranspose_MPIAIJ does not work

Thomas Witkowski thomas.witkowski at tu-dresden.de
Mon Oct 29 02:36:24 CDT 2012


Thanks for the very fast fix!

Thomas

Am 29.10.2012 08:34, schrieb Jed Brown:
> Sorry about that, this is what I _thought_ I had typed.
>
> https://bitbucket.org/petsc/petsc-dev/changeset/6b46d15782069ee
>
> On Mon, Oct 29, 2012 at 12:24 AM, Thomas Witkowski 
> <thomas.witkowski at tu-dresden.de 
> <mailto:thomas.witkowski at tu-dresden.de>> wrote:
>
>     Hi,
>
>     you made a workaround for the MatTranspose problem with one-side
>     communications:
>
>     https://bitbucket.org/petsc/petsc-dev/changeset/43e28ef1a420
>
>     My code now crashes on MatTranspose in line 2147 of
>     src/mat/impls/aij/mpi/mpiaij.c, as the matrix object is not
>     created but MatSetOption is executed on this matrix.
>
>     Best regards,
>
>     Thomas
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20121029/78ccbb11/attachment.html>


More information about the petsc-dev mailing list