inria-00451471, version 2
Optimizing MPI Communication within large Multicore nodes with Kernel assistance
Stéphanie Moreaud
a, 1, 2Brice Goglin 1, 2David Goodell 3Raymond Namyst 1, 2
Workshop on Communication Architecture for Clusters, held in conjunction with IPDPS 2010 (2010) 7 p.
Résumé : As the number of cores per node increases in modern clusters, intra-node communication efficiency becomes critical to application performance. We present a study of the traditional double-copy model in MPICH2 and a kernel-assisted single-copy strategy with KNEM on different shared-memory hosts with up to 96 cores. We show that KNEM suffers less from process placement on these complex architectures. It improves throughput up to a factor of 2 for large messages for both point-to-point and collective operations, and significantly improves NPB execution time. We detail when to switch from one strategy to the other depending on the communication pattern and we show that \ioat copy offload only appears to be an interesting solution for older architectures.
- a – INRIA
- 1 : Laboratoire Bordelais de Recherche en Informatique (LaBRI)
- CNRS : UMR5800 – Université Sciences et Technologies - Bordeaux I – Ecole Nationale Supérieure d'Electronique, Informatique et Radiocommunications de Bordeaux – Université Victor Segalen - Bordeaux II
- 2 : RUNTIME (INRIA Bordeaux - Sud-Ouest)
- INRIA – CNRS : UMR5800 – Université Sciences et Technologies - Bordeaux I – Ecole Nationale Supérieure d'Electronique, Informatique et Radiocommunications de Bordeaux
- 3 : Mathematics and Computer Science Division [ANL] (MCS)
- Argonne National Laboratory
- Domaine : Informatique/Système d'exploitation
- Versions disponibles : v1 (29-01-2010) v2 (02-02-2010)
- inria-00451471, version 2
- http://hal.inria.fr/inria-00451471
- oai:hal.inria.fr:inria-00451471
- Contributeur : Brice Goglin
- Soumis le : Mardi 2 Février 2010, 16:09:32
- Dernière modification le : Lundi 14 Juin 2010, 08:24:50
Documents associés
Exporter