Gotcha. Thanks -- I'll do a quick run-through of the code and ensure that we only use ompi_comm_remote_site() in the appropriate places.

On Aug 17, 2005, at 11:47 AM, Edgar Gabriel wrote:

Jeff,

as far as I know nothing has changed in the communicator setup since
quite a while, the remote group is still set to the local group for
intra-communicators. However, ompi_comm_remote_size has a check whether
we have an inter-comm, and returns else zero (which is in theory the
correct answer)....

Thanks
Edgar

Jeff Squyres wrote:
I thought that we set the remote group equal to the local group in
communicator setup such that ompi_comm_remote_size() was *always* safe
to call...?

This no longer seems to be true -- while debugging the IMB (previously
the pallas benchmarks), I found that calling
ompi_comm_remote_size(comm) would return 0 and ompi_comm_size(comm)
would return 2 for a simple intracommunicator.  This was with:

        mpirun -np 2 IMB-MPI1 Reduce

Has something changed in communicator setup?



--
======================================================================
Dr.-Ing. Edgar Gabriel
Clusters and Distributed Units
High Performance Computing Center Stuttgart (HLRS)
University of Stuttgart
Tel: +49 711 685 8039                http://www.hlrs.de/people/gabriel
Fax: +49 711 678 7626                e-mail:gabr...@hlrs.de
======================================================================
_______________________________________________
devel mailing list
de...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/devel


--
{+} Jeff Squyres
{+} The Open MPI Project
{+} http://www.open-mpi.org/

Reply via email to