Advertisement

Correction of Monitor Intrusion for Testing Nondeterministic MPI-Programs

  • D. Kranzlmüller
  • Ch. Schaubschläger
  • J. Chassin de Kergommeaux
Conference paper
  • 46 Downloads
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1685)

Abstract

Software analysis tools apply monitors to retrieve state information about program executions. Unfortunately such observation introduces the probe effect, which means that analysis data are influenced by the monitor’s requirements to computing time and space. Additionally, nondeterministic parallel programs may yield different execution patterns due to alterations of event ordering at race conditions. In order to correct these intrusions, two activities are carried out. Firstly, the monitor overhead is removed by recalculating event occurrence times based on measurements of the occurred delay. Secondly, event manipulation and program replay is applied at places, where reordering of events has occurred. The resulting data describes the program’s execution without monitoring.

References

  1. [1]
    J. Gait: A Probe Effect in Concurrent Programs. Software-Practise and Experience, Vol. 16(23), pp. 225–233 (March 1986).Google Scholar
  2. [2]
    D. Kranzlmüller, S. Grabner, J. Volkert: Debugging with the MAD Environment. Parallel Computing, Vol. 23, No. 1-2, pp. 199–217 (Apr. 1997).Google Scholar
  3. [3]
    L. Lamport: Time, Clocks, and the Ordering of Events in a Distributed System. Comm. ACM, pp. 558–565 (July 1978).Google Scholar
  4. [4]
    E. Maillet, C. Tron: On Efficiently Implementing Global Time for Performance Evaluation on Multiprocessor Systems. Journal of Parallel and Distributed Computing, Vol. 28, pp. 84–93 (July 1995).Google Scholar
  5. [5]
    A. Malony, D. Reed: Models for performance perturbation analysis. Proc. Workshop Parallel Distributed Debugging, ACMSIGPLAN/SIGOPS and Office of Naval Research, (May 1991).Google Scholar
  6. [6]
    Message Passing Interface Forum: MPI: A Message-Passing Interface Standard-Version 1.1. http://www.mcs.anl.gov/mpi/ (June 1995).
  7. [7]
    R. Reussner, P. Sanders, L. Prechelt, M. Müller: SKaMPI: A Detailed, Accurate MPI Benchmark. Proc. 5th European PVM/MPI Users’Group Meeting, Springer, Lecture Notes in Computer Science, Vol. 1497, Liverpool, UK, pp. 52–59 (Sept. 1998).Google Scholar
  8. [8]
    M. Ronsse, D. Kranzlmüller: RoltMP-Replay of Lamport Timestamps for Message Passing Systems. Proc. 6th EUROMICRO Workshop on Parallel and Distributed Processing, University of Madrid, Spain, pp. 87–93, (Jan. 1998).Google Scholar
  9. [9]
    F. Teodorescu, J. Chassin de Kergommeaux: On Correcting the Intrusion of Tracing Non-deterministic Programs by Software. Proc. EUROPAR’97 Parallel Processing, 3rd Intl. Euro-Par Conference, Springer, Lecture Notes in Computer Science, Vol. 1300, Passau, Germany, pp. 94–101 (Aug. 1997).Google Scholar
  10. [10]
    W. Wu, R. Gupta, M. Spezialetti: Experimental Evaluation of On-line Techniques for Removing Monitoring Intrusion. Proc. of SPDT’98, SIGMETRICS Symposium on Parallel and Distributed Tools, Welches, Oregon, pp. 30–39, (Aug. 1998).Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1999

Authors and Affiliations

  • D. Kranzlmüller
    • 1
  • Ch. Schaubschläger
    • 1
  • J. Chassin de Kergommeaux
    • 2
  1. 1.GUP Linz, Joh. Kepler University LinzAustria
  2. 2.ID-IMAGGrenobleFrance

Personalised recommendations