DMCA
Intra node parallelization of MPI programs with OpenMP (1998)
| Citations: | 3 - 0 self |
BibTeX
@TECHREPORT{Cappello98intranode,
author = {Franck Cappello and Olivier Richard},
title = {Intra node parallelization of MPI programs with OpenMP},
institution = {},
year = {1998}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
The availability of multiprocessors and high performance networks offer the opportunity to construct CLUMPs (Cluster of Multiprocessors) and use them as paxallel computing platforms. The main distinctive feature of the CLUMP axchitecture over the usual paxallel computers is its hybrid memory model (message passing between the nodes and shaxed memory inside the nodes). Some of the primaxy issues to address for the CLUMP axe: 1) to be able to execute the existing programs with few modifications 2) to provide some programming models coherent with the performance hieraxchy of the data movements inside the CLUMP 3) to limit the effort of the programmer while ensuring the portability of the codes on a wide vaxiety of CLUMP configurations. We investigate an approach based on the MPI and OpenMP standaxds. The approach consists in the intra-node paxallelization of the MPI programs with an OpenMP directire based paxallel compiler. The paper presents a detailed study of the approach in the context of the biprocessor PC CLUMPs. It provides three contributions. First, it evaluates the ability of biprocessor PCs to effectively provide a speed up over single processor PCs in the context of shaxed memory paxallel programs. Second, it investigates the method to transform MPI paxallel programs in order to execute them on a CLUMP. Third, it presents the performance evaluation of this method applied on the NAS paxallel benchmaxks executed on a cluster of biprocessor PCs.
Keyphrases
mpi program intra node parallelization biprocessor pc biprocessor pc clump clump axe openmp standaxds openmp directire wide vaxiety message passing performance hieraxchy usual paxallel computer clump axchitecture data movement intra-node paxallelization main distinctive feature mpi paxallel program paxallel compiler hybrid memory model detailed study performance evaluation shaxed memory paxallel program primaxy issue high performance network na paxallel benchmaxks clump configuration single processor pc shaxed memory programming model coherent
