Advertisement

SKaMPI — Towards Version 5

  • Werner Augustin
  • Michael Haller
  • Marc-Oliver Straub
  • Thomas Worsch
Conference paper

Abstract

SKaMPI is now an established benchmark for MPI implementations. The development of SKaMPI-5 strives for improvements in several directions: (i) extension of the benchmark to cover more functionality of MPI, (ii) construction of a collection of collective algorithm kernels which are not supported by core MPI collective operations. (iii) a redesign of the SKaMPI benchmark allowing it to be extended more easily (thus matching requests from SKaMPI users).

In the present paper we give an overview of the extension of SKaMPI for the evaluation of virtual topologies, describe the foundations of new algorithms for fast all-to-all communication specifically tailored for the case of differing message sizes, and give a first impression of what SKaMPI-5 will look like, for which we now have a prototype running.

Keywords

Message Passing Interface Configuration File Message Size Virtual Topology Collective Communication 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    W. Augustin and T. Worsch. Usefulness and usage of SKaMPI-bench. In J. Dongarra, D. Laforenza, and S. Orlando, eds., Proc. 10th Euro PVM/MPI, LNCS 2840, pages 63–70. Springer, 2003.Google Scholar
  2. 2.
    Message Passing Interface Forum. MPI: A message passing interface standard, 1995.Google Scholar
  3. 3.
    M. Haller and T. Worsch. SKaMPI — including more complex communication patterns. In E. Krause, W. Jäger, and M. Resch, eds., High Performance Computing in Science and Engineering’ 03, pages 455–466. Springer, 2003.Google Scholar
  4. 4.
    M. Kaufmann, J.F. Sibeyn, and T. Suel. Beyond the bisection bound: fast ranking and counting on meshes. In P.G. Spirakis, ed., Proc. 3rd European Symposium on Algorithms, pages 75–88, LNCS 979. Springer, 1995.Google Scholar
  5. 5.
    R.H. Reussner, P. Sanders, and J.L. Träff. SKaMPI: a comprehensive benchmark for public benchmarking of MPI. Scientific Programming, 10(1):55–65, 2002.Google Scholar
  6. 6.
    L.G. Valiant and G. J. Brebner. Universal schemes for parallel communication. In Conference proceedings of the Thirteenth Annual ACM Symposium on Theory of Computing: papers presented at the Symposium, Milwaukee, Wisconsin, May 11–13, 1981, pages 263–277, New York, NY, USA, 1981. ACM Press.Google Scholar
  7. 7.
    T. Worsch, R.H. Reussner, and W. Augustin. Benchmarking collective opera tions with SKaMPI. In E. Krause and W. Jäger, eds., High Performance Computing in Science and Engineering’ 02, pages 491–502. Springer, 2002.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2005

Authors and Affiliations

  • Werner Augustin
    • 1
  • Michael Haller
    • 1
  • Marc-Oliver Straub
    • 1
  • Thomas Worsch
    • 1
  1. 1.LIINUniversität KarlsruheGermany

Personalised recommendations