The Architecture and Performance of WMPI II

  • Anders Lyhne Christensen
  • João Brito
  • João Gabriel Silva
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3241)


WMPI II is the only commercial implementation of MPI 2.0 that runs on both Windows and Linux clusters. It evolved from the first ever Windows version of MPI, then a port of MPICH, but is now fully built from its own code base. It supports both 32 and 64 bit versions and mixed clusters of Windows and Linux nodes. This paper describes the main design decisions and the multithreaded, non-polling architecture of WMPI II. Experimental results show that, although WMPI II has figures comparable to MPICH and LAM for latency and bandwidth, most application benchmarks perform significantly better when running on top of WMPI II.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [MAR]
    Marinho, J., Silva, J.G.: WMPI-Message Passing Interface for Win32 Clusters. In: Proc. EuroPVM/MPI 1998, Liverpool, September 1998, Springer, Heidelberg (1998)Google Scholar
  2. [MPIR1]
    Snir, M., et al.: MPI - The Complete Reference, vol. 1. MIT Press, Cambridge (1998)Google Scholar
  3. [MPIR2]
    Gropp, W., et al.: MPI - The Complete Reference, vol. 2. MIT Press, Cambridge (1998)Google Scholar
  4. [MPIS1]
    MPI Forum, MPI: A Message-Passing Interface Standard (1994)Google Scholar
  5. [MPIS2]
    MPI Forum, MPI-2: Extensions to the Message-Passing Interface (1997)Google Scholar
  6. [LAMW]
    The LAM group maintains a list of available MPI implementations and their features on
  7. [HPI]
    Group, W., et al.: High-performance, portable implementation of the Message Passing Interface Standard. Journal of Parallel Computing 22(6) (1996)Google Scholar
  8. [LAM]
    Burns, G., et al.: LAM: An Open Cluster Environment for MPI. In: Proceedings of Supercomputing Symposium (1994)Google Scholar
  9. [CHA]
    Gropp, W., Lusk, E.: MPICH working note: Creating a new MPICH device using the channel interface, Technical Report ANL/MCSTM- 213, Argonne National Laboratory (1995)Google Scholar
  10. [OMP]
    OpenMP: An Industry-Standard API for Shared-Memory Programming. IEEE Computational Science & Engineering 5(1) (January/March 1998) Google Scholar
  11. [PAL]
    Pallas MPI Benchmarks - PMB, Part MPI-1, Pallas GmbH (2000) Google Scholar
  12. [NAS]
    Bailey, D.H., et al.: The NAS Parallel Benchmarks. International Journal of Supercomputer Applications (1991)Google Scholar
  13. [LIN]
    Dongarra, J.J.: Performance of Various Computers Using Standard Linear Equations Software, University of Tennesse (1995)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Anders Lyhne Christensen
    • 1
  • João Brito
    • 1
  • João Gabriel Silva
    • 2
  1. 1.Critical Software SA 
  2. 2.DEI/CISUCUniversity of CoimbraPortugal

Personalised recommendations