Performance Modeling Based on PVM

  • Hermann Mierendorff
  • Helmut Schwamborn
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1697)


For performance modeling of message passing programs using PVM, we consider a hybrid method where large numerical parts of the program are replaced byde lays and the remaining control structure is translated into the input language of a simulator. For simulation, we use again a PVM-like platform. The present paper deals mainlyw ith special aspects of the PVM layer used for simulation. As a case study, a subset of PVM was implemented on top of Modarch which is a commercial tool for simulating parallel systems. It was used here as a simulator for performance models of parallel Fortran programs. Some experiences with this implementation are discussed. This wayo f performance evaluation is compared with a method where the model is written in the same language as the original program and the simulator consists of a set of routines complementing the original PVM layer.


Automatic performance modeling parallel programming 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    A. Geist, A. Beguelin, J. Dongarra, W. Jiang, R. Manchek, and V. Sunderam, PVM: Parallel Virtual Machine-A User’s Guide and Tutorial for Networked Parallel Computing, (MIT Press, Cambridge, MA, 1994).Google Scholar
  2. 2.
    H. Mierendorff and H. Schwamborn, Generation of Performance Models, in D’Hollander, Joubert, Peters, Trottenberg, Parallel Computing: Fundamentals, Applications and New Directions (Elsevier Science, Amsterdam, 1998) 693–696.Google Scholar
  3. 3.
    H. Mierendorff and H. Schwamborn, Automatic Model Generation for Performance Estimation of Parallel Programs, (accepted for Parallel Computing).Google Scholar
  4. 4.
    E. van de Velde, Concurrent Scientific Computing (Springer-Verlag, New York, 1992).Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1999

Authors and Affiliations

  • Hermann Mierendorff
    • 1
  • Helmut Schwamborn
    • 1
  1. 1.Institute for Algorithms and Scientific Computing (SCAI)GMD - German National Research Center for Information TechnologySankt AugustinGermany

Personalised recommendations