Comparison of MPI benchmark programs on an SGI altix ccNUMA shared memory machine

Simple item page
dc.contributor.authorAbdul Hamid, N.
dc.contributor.authorCoddington, P.
dc.contributor.authorVaughan, F.
dc.contributor.conferenceInternational Parallel and Distributed Processing Symposium (20th : 2006 : Rhodes Island, Greece)
dc.contributor.editorOuid Khaoua, M.
dc.contributor.editorMin, G.
dc.date.issued2006
dc.description.abstractThe results produced by five different MPI benchmark programs on an SGI Altix 3700 are analyzed and compared. There are significant differences in the results for some MPI operations. We investigate the reasons for these discrepancies, which are due to differences in the measurement techniques, implementation details and default configurations of the different benchmarks. The variation in results on the Altix are generally much greater than on a distributed memory machine, due primarily to the ccNUMA architecture and the importance of cache effects, as well as some implementation details of the SGI MPI libraries.
dc.identifier.citationParallel and Distributed Processing Symposium, IPDPS 20th International, 25-29 April, 2006: 8p.
dc.identifier.doi10.1109/IPDPS.2006.1639631
dc.identifier.isbn1424400546
dc.identifier.orcidCoddington, P. [0000-0003-1336-9686]
dc.identifier.urihttp://hdl.handle.net/2440/36008
dc.language.isoen
dc.publisherIEEE
dc.publisher.placeUSA
dc.source.urihttps://doi.org/10.1109/ipdps.2006.1639631
dc.titleComparison of MPI benchmark programs on an SGI altix ccNUMA shared memory machine
dc.typeConference paper
pubs.publication-statusPublished

Files

Collections

Computer Science publications