[petsc-users] Destroy ghosted vector that was used to duplicate

[John Mousel]

I was looking at the number of persistent sends and receives associated
with scatters for ghosted vectors using the Totalview Message Queue Graph.
These were growing because I was not destroying a duplication of a
duplication of a vector. I was not observing an incorrect number of
communicators, which was just a single duplication of MPI_COMM_WORLD.

John


On Wed, Jun 19, 2013 at 5:09 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> John Mousel <john.mousel at gmail.com> writes:
>
> > Looking at the message queue graph in Totalview, it seems that I'm
> running
> > up the number of persistent communications when I try to resize a vector
> by
> > duplicating, copying, destroying, and calling VecCreateGhost.
>
> I don't know how you're measuring that.  PETSc does not create new
> communicators per object.  We create one PETSc comm per user comm, and
> all the objects share that.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20130619/a7e37b84/attachment.html>