On Dec 19, 2011, at 12:06 AM, Dave Nystrom wrote:
> Several of the external packages that have been interfaced with petsc require
> mpi. Is there a way to build a serial version of petsc with these packages
> and use stub routines for mpi?
MUMPS can be built without MPI using petsc-dev but a
On Mon, 19 Dec 2011, Barry Smith wrote:
>
> On Dec 19, 2011, at 12:06 AM, Dave Nystrom wrote:
>
> > Several of the external packages that have been interfaced with petsc
> > require
> > mpi. Is there a way to build a serial version of petsc with these packages
> > and use stub routines for mpi
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/20111219/4df3e
> consecutive integers are handled well.
>
> So far, I cannot find any free implementations.
>
Do you have a reference showing that the constants are decent?
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111219/1eebd295/attachment.html>
heir
experiments lead.
-- Norbert Wiener
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111219/5786fd2a/attachment.html>
On Dec 19, 2011, at 9:30 PM, Matthew Knepley wrote:
> I looked at the new checkins for fast MatMatSym. It looks to me like you want
> a data
> structure that has
>
> a) O(N) space, where N is the number of entries
>
> b) fast insert
>
> c) fast traversal
>
> It seems like the best dat
Note that this same operation (merging collections of sorted integers) is
also needed for factorization so having a solid toolkit of different algorithms
that perform this well for different size problems and different levels of
sparsity would be a useful thing in PETSc.
Barry
On Dec 1
