Have you checked where the procs are actually bound? I’m wondering if you don’t wind up somehow with a proc bound to cores on different sockets
> On Jul 6, 2015, at 7:06 AM, Saliya Ekanayake <esal...@gmail.com> wrote: > > Just checking if anyone has experienced a similar situation or has any > pointers to understand this. > > Thank you > Saliya > > On Jul 1, 2015 9:27 PM, "Saliya Ekanayake" <esal...@gmail.com > <mailto:esal...@gmail.com>> wrote: > Hi, > > I am getting strange performance results for allgatherv operation for the > same number of procs and data, but with varying binding width. For example > here are two cases with about 180x difference in performance. > > Each machine has 4 sockets each with 6 cores totaling 24 cores per node > (topology attached). > > Case 1 > ---- > 12 procs per node each bound to 1 core times 30 nodes --> 1929 ms > > Case 2 > ---- > 12 procs per node each bound to 2 cores times 30 nodes --> 357209 ms > > > Another set of variations for 2 procs per node and 4 procs per node is given > below in the chart. Is such variation expected with binding width? I am a bit > puzzled and would appreciate any help to understand this. > > <image.png> > > Thank you, > Saliya > > -- > Saliya Ekanayake > Ph.D. Candidate | Research Assistant > School of Informatics and Computing | Digital Science Center > Indiana University, Bloomington > Cell 812-391-4914 <tel:812-391-4914> > http://saliya.org > <http://saliya.org/>_______________________________________________ > users mailing list > us...@open-mpi.org <mailto:us...@open-mpi.org> > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users > <http://www.open-mpi.org/mailman/listinfo.cgi/users> > Link to this post: > http://www.open-mpi.org/community/lists/users/2015/07/27258.php > <http://www.open-mpi.org/community/lists/users/2015/07/27258.php>