[ET Trac] [Einstein Toolkit] #2187: comet files in simfactory use in MPI rank per node

Einstein Toolkit trac-noreply at einsteintoolkit.org
Wed Aug 22 10:09:32 CDT 2018


#2187: comet files in simfactory use in MPI rank per node
---------------------------------+-------------------------
 Reporter:  Roland Haas          |       Type:  enhancement
   Status:  new                  |   Priority:  minor
Milestone:                       |  Component:  SimFactory
  Version:  development version  |   Keywords:
---------------------------------+-------------------------
 The current
 (https://bitbucket.org/simfactory/simfactory2/src/master/mdb/machines/comet.ini)
 uses 1 MPI rank per node:
 {{{
 max-num-threads = 24
 num-threads     = 24
 }}}

 This is usually not the best way to set things up, I would eg have
 expected that the default choice would be something like 1 MPI rank per
 NUMA domain.
 Given that, unless limited by communication overhead, we seem to obtain
 fastest per-node performance when using only MPI and no OpenMP (about a
 factor of 50% speedup on my 12 core workstation with 2 NUMA domains) if
 anyone is using Comet for production work and wants to contribute their
 machine description file that would be great.

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/2187>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit


More information about the Trac mailing list