Optimizing MPI Communication within large Multicore nodes with Kernel assistance

Stéphanie Moreaud (1, 2) , Brice Goglin (1, 2) , David Goodell (3) , Raymond Namyst (1, 2)

1 LaBRI - Laboratoire Bordelais de Recherche en Informatique
2 RUNTIME - Efficient runtime systems for parallel architectures
3 MCS - Mathematics and Computer Science Division [ANL]

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.

Domaines: Système d'exploitation [cs.OS]

Stéphanie Moreaud, Brice Goglin, David Goodell, Raymond Namyst. Optimizing MPI Communication within large Multicore nodes with Kernel assistance. Workshop on Communication Architecture for Clusters, held in conjunction with IPDPS 2010, Apr 2010, Atlanta, United States. 7 p., ⟨10.1109/IPDPSW.2010.5470849⟩. ⟨inria-00451471⟩