Hybrid MPI/StarSs - a case study

Jose Gracia; Christoph Niethammer; Manuel Hasert; Steffen Brinkmann; Rainer Keller; Colin W. Glass

Hybrid MPI/StarSs - a case study

Distributed, Parallel, and Cluster Computing 2012-04-19 v1

Authors: Jose Gracia , Christoph Niethammer , Manuel Hasert , Steffen Brinkmann , Rainer Keller , Colin W. Glass

Abstract

Hybrid parallel programming models combining distributed and shared memory paradigms are well established in high-performance computing. The classical prototype of hybrid programming in HPC is MPI/OpenMP, but many other combinations are being investigated. Recently, the data-dependency driven, task parallel model for shared memory parallelisation named StarSs has been suggested for usage in combination with MPI. In this paper we apply hybrid MPI/StarSs to a Lattice-Boltzmann code. In particular, we present the hybrid programming model, the benefits we expect, the challenges in porting, and finally a comparison of the performance of MPI/StarSs hybrid, MPI/OpenMP hybrid and the original MPI-only versions of the same code.

Keywords

parallel programming

Cite

@article{arxiv.1204.4086,
  title  = {Hybrid MPI/StarSs - a case study},
  author = {Jose Gracia and Christoph Niethammer and Manuel Hasert and Steffen Brinkmann and Rainer Keller and Colin W. Glass},
  journal= {arXiv preprint arXiv:1204.4086},
  year   = {2012}
}

Comments

8 pages, accepted for publication in ISPA 2012

Hybrid MPI/StarSs - a case study

Abstract

Keywords

Cite

Comments

Related papers