Advertisement

PerPreT — A performance prediction tool for massively parallel systems

  • Jürgen Brehm
  • Manish Madhukar
  • Evgenia Smirni
  • Larry Dowdy
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 977)

Abstract

Today's massively parallel machines are typically message passing systems consisting of hundreds or thousands of processors. Implementing parallel applications efficiently in this environment is a challenging task. The Performance Prediction Tool (PerPreT) presented in this paper is useful for system designers and application developers. The system designers can use the tool to examine the effects of changes of architectural parameters on parallel applications (e.g., reduction of setup time, increase of link bandwidth, faster execution units). Application developers are interested in a fast evaluation of different parallelization strategies of their codes. PerPreT uses a relatively simple analytical model to predict speedup, execution time, computation time, and communication time for a parametrized application. Especially for large numbers of processors, PerPreT's analytical model is preferable to traditional models (e.g., Markov based approaches such as queueing and Petri net models). The applications are modelled through parameterized formulae for communication and computation. The parameters used by PerPreT include the problem size and the number of processors used to execute the program. The target systems are described by architectural parameters (e.g., setup times for communication, link bandwidth, and sustained computing performance per node).

Keywords

workload modeling performance evaluation performance prediction 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [Har94]
    Günter Haring, Harald Wabnig: PAPS — The Parallel Program Performance Prediction Toolset, Proceedings of the 7th Int. Conf. on Modelling Techniques and Tools for Computer Performance Evaluation, LNCS 794, pp. 284–304, Springer Verlag, 1994.Google Scholar
  2. [Har95]
    Günter Haring, Gabriele Kostis: Workload Modeling for Parallel Processing Systems, Proceedings of the 3rd International Workshop on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS 95), IEEE Computer Society Press, pp. 8–12, Durham, NC 1995.Google Scholar
  3. [Laz84]
    E. D. Lazowska et. al.: Quantitative System Performance: Computer System Analysis using Queueing Network Models, Englewood Cliffs, NJ, Prentice Hall, 1984.Google Scholar
  4. [LOOP94]
    J. Brehm et. al.: A Multiprocessor Communication Benchmark, User's Guide and Reference Manual, Public Report of the ESPRIT III Benchmarking Project, 1994.Google Scholar
  5. [Meh94]
    Pankaj Mehra et. al.: A Comparison of Two Model-Based Performance Prediction Techniques for Message Passing Parallel Programs, Proceedings of the ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, Nashville, TN, May 1994.Google Scholar
  6. [NAS93]
    D.H. Bailay et. al.: NAS Parallel Benchmarks Results, Parallel and Distributed Technology, Vol. 1, IEEE, February 1993.Google Scholar
  7. [Ser93]
    Maria Calzarossa, Giuseppe Serazzi: Workload Characterization — A Survey, Proceedings of the IEEE, 81(8), pp. 1136–1150, August 1993.CrossRefGoogle Scholar
  8. [Par94]
    D. Walker et al.: Public International Benchmarks for Parallel Computers, Report of the ParkBench Committee, available on www: http://www.epm.ornl.gov/∼walker/report.html.Google Scholar
  9. [PICL90]
    P.H. Worley et. al.: PICL — A Portable Instrumented Communication Library, Technical Report, ORNL/TM-1130, Oak Ridge National Laboratory, Oak Ridge, July 1990.Google Scholar
  10. [Smi95]
    E. Smirni et. al.: Thread Placement on the Intel Paragon: Modeling and Experimenation, Proceedings of the 3rd International Workshop on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS 95), IEEE Computer Society Press, pp. 226–231, Durham, NC 1995.Google Scholar
  11. [Tho86]
    A. Thomasian, Paul F. Bay: Analytic Queueing Network Models for Parallel Processing of Task Systems, IEEE Transaction on Computers, Vol. C-35, No.12, December 1986.Google Scholar
  12. [Tri82]
    K. S. Trivedi, P. Heidelberger: Queueing Network Models for Parallel Processing with Asynchronous Tasks, IEEE Transactions on Computers, Vol C-32, pp. 15–31, January 1982.Google Scholar
  13. [Wab94]
    Harald Wabnig, Günter Haring: Performance Prediction of Parallel Systems with Scalable Specifications — Methodology and Case Study, Proceedings of the ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, pp. 288–289, Nashville, TN, 1994.Google Scholar
  14. [Wor94]
    P. H. Worley, I. T. Foster. Parallel Spectral Transform Shallow Water Model: A Runtime-Tunable Parallel Benchmark Code, Proceedings of the SHPCC'94, IEEE Computer Society, pp. 207–214, 1994Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1995

Authors and Affiliations

  • Jürgen Brehm
    • 1
  • Manish Madhukar
    • 2
  • Evgenia Smirni
    • 2
  • Larry Dowdy
    • 2
  1. 1.Institut für Rechnerstrukturen und BetriebssystemeUniversität HannoverHannover
  2. 2.Department of Computer ScienceVanderbilt UniversityNashville

Personalised recommendations