The scalable process topology interface of MPI 2.2

Authors

  • Torsten Hoefler,

    1. University of Illinois at Urbana-Champaign, 1205 W. Clark St. Urbana, IL 61801, U.S.A.
    Search for more papers by this author
    • The work of this author was done mostly at Indiana University, IN, U.S.A.

  • Rolf Rabenseifner,

    1. HLRS, University of Stuttgart, D-70550 Stuttgart, Germany
    Search for more papers by this author
  • Hubert Ritzdorf,

    1. HPCE, NEC Deutschland GmbH, Hansaallee 101, 40549 Düsseldorf, Germany
    Search for more papers by this author
    • The work of these authors was done mostly at NEC Laboratories Europe, St. Augustin, Germany.

  • Bronis R. de Supinski,

    1. Lawrence Livermore National Laboratory, Box 808, L-557, Livermore, CA 94551-0808, U.S.A.
    Search for more papers by this author
  • Rajeev Thakur,

    1. Mathematics and Computer Science Division, Argonne National Laboratory, 9700 S. Cass Avenue, Argonne, IL 60439, U.S.A.
    Search for more papers by this author
  • Jesper Larsson Träff

    Corresponding author
    1. Department of Scientific Computing, University of Vienna, Nordbergstrasse 15/3C, A-1090 Vienna, Austria
    • Department of Scientific Computing, University of Vienna, Nordbergstrasse 15/3C, A-1090 Vienna, Austria
    Search for more papers by this author
    • The work of these authors was done mostly at NEC Laboratories Europe, St. Augustin, Germany.


Abstract

The Message-passing Interface (MPI) standard provides basic means for adaptations of the mapping of MPI process ranks to processing elements to better match the communication characteristics of applications to the capabilities of the underlying systems. The MPI process topology mechanism enables the MPI implementation to rerank processes by creating a new communicator that reflects user-supplied information about the application communication pattern. With the newly released MPI 2.2 version of the MPI standard, the process topology mechanism has been enhanced with new interfaces for scalable and informative user-specification of communication patterns. Applications with relatively static communication patterns are encouraged to take advantage of the mechanism whenever convenient by specifying their communication pattern to the MPI library. Reference implementations of the new mechanism can be expected to be readily available (and come at essentially no cost), but non-trivial implementations pose challenging problems for the MPI implementer. This paper is first and foremost addressed to application programmers wanting to use the new process topology interfaces. It explains the use and the motivation for the enhanced interfaces and the advantages gained even with a straightforward implementation. For the MPI implementer, the paper summarizes the main issues in the efficient implementation of the interface and explains the optimization problems that need to be (approximately) solved by a good MPI library. Copyright © 2010 John Wiley & Sons, Ltd.

Ancillary