[petsc-dev] What do people want to have working before a petsc-3.2 release?

[Jed Brown]

On Wed, Dec 22, 2010 at 01:00, Sean Farley <sean at mcs.anl.gov> wrote:
>
> Fair enough but I was just bitten by the fact that SuperLU_DIST requires
> ParMetis :-(
>

Would SuperLU_Dist be 64-bit clean if not for ParMetis?  What did they do to
ParMetis to get it to work?

It appears that SuperLU_Dist only uses ParMETIS_V3_NodeND, which is one of
the functions wrapped by Scotch.  That API still hard-codes int, but the
interface is thin enough that it might be trivial to make it available with
64-bit ints.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20101222/bd998de3/attachment.html>