A Benchmark for MPI Derived Datatypes

  • Ralf Reussner
  • Jesper Larsson Träff
  • Gunnar Hunzelmann
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1908)

Abstract

We present an extension of the SKaMPI benchmark for MPI implementations to cover the derived datatype mechanism of MPI. All MPI constructors for derived datatypes are covered by the benchmark, and varied along different dimensions. This is controlled bya set of predened patterns which can be instantiated bypa rameters given bythe user in a configurationsle. We classifythe patterns intofixed types, dynamic types, nested types, and special types. We show results from the SKaMPI ping-pong measurement with the fixed and special types on three platforms: CrayT3E/900, IBM RS 6000SP, NEC SX-5. The machines show quite some difference in handling datatypes, with typically a significant penaltyfor nested types for the Cray (up to a factor of 16) and the IBM (up to a factor of 8), whereas the NEC treats these types very uniformly (overhead of between 2 and 4). Such results illustrate the need for a systematic datatype benchmark to help the MPI programmer select the most efficient data representation for a particular machine.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    D. Bailey, E. Barszcz, J. Barton, D. Browning, and R. Carter. The NAS parallel benchmarks. Technical Report RNR-94-007, RNR, 1994.Google Scholar
  2. [2]
    W. Gropp, S. Huss-Lederman, A. Lumsdaine, E. Lusk, B. Nitzberg, W. Saphir, and M. Snir. MPI-The Complete Reference, volume 2, The MPI Extensions. MIT Press, 1998.Google Scholar
  3. [3]
    W. Gropp and E. Lusk. Reproducible measurements of MPI performance characteristics. In Recent Advances in Parallel Virtual Machine and Message Passing Interface. 6th European PVM/MPI Users’ Group Meeting, volume 1697 of Lecture Notes in Computer Science, pages 11–18, 1999.CrossRefGoogle Scholar
  4. [4]
    W. Gropp, E. Lusk, N. Doss, and A. Skjellum. A high-performance, portable imlementation of the MPI message passing interface standard. Parallel Computing, 22(6):789–828, 1996.MATHCrossRefGoogle Scholar
  5. [5]
    Rolf Hempel. Basic message passing benchmarks, methodology and pitfalls, September 1999. Presented at the SPEC Workshop, slides available at http://www.hlrs.de/mpi/b_eff/hempel_wuppertal.ppt
  6. [6]
    Parkbench Committee. Public International Benchmarks for Parallel Computers. Scientific Programming, 3(2):101–146, 1994. Report 1.Google Scholar
  7. [7]
    R. Reussner, P. Sanders, L. Prechelt, and M. Müller. SKaMPI: A detailed, accurate MPI benchmark. In Recent Advances in Parallel Virtual Machine and Message Passing Interface. 5th European PVM/MPI Users’ Group Meeting, volume 1497 of Lecture Notes in Computer Science, pages 52–59, 1998.CrossRefGoogle Scholar
  8. [8]
    M. Snir, S. Otto, S. Huss-Lederman, D. Walker, and J. Dongarra. MPI-The Complete Reference, volume 1, The MPI Core. MIT Press, second edition, 1998.Google Scholar
  9. [9]
    J. L. Träff, R. Hempel, H. Ritzdorf, and F. Zimmermann. Flattening on the fly: efficient handling of MPI derived datatypes. In Recent Advances in Parallel Virtual Machine and Message Passing Interface. 6th European PVM/MPI Users’ Group Meeting, volume 1697 of Lecture Notes in Computer Science, pages 109–116, 1999.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2000

Authors and Affiliations

  • Ralf Reussner
    • 1
  • Jesper Larsson Träff
    • 2
  • Gunnar Hunzelmann
    • 1
  1. 1.LIINUniversität KarlsruheKarlsruheGermany
  2. 2.C&C Research LaboratoriesNEC Europe Ltd.Sankt AugustinGermany

Personalised recommendations