Dr. Richard S. Miller
Assistant Professor
Department of Mechanical Engineering
Clemson University
Last updated: 10/17/2000

The following pages contain a brief introduction to parallel processing focusing primarily on using the Message Passing Interface (MPI) communication subroutines for applications on either shared or distributed memory parallel computer architectures. The material presented is only meant to be an introduction to basic concepts of parallel computers and the MPI communication routines. It is not meant to be a stand-alone document with which the reader can become proficient in MPI programming; recommended references to proficient MPI programming are provided at the bottom of this page. The writing of this document was motivated in part by the construction of a beowulf cluster of dual processor PCs by the Department of Mechanical Engineering; however, the concepts discussed in what follows are not specific to beowulf clusters. The document covers subjects that in the author's experience have greatly benefited incoming graduate students in the process of being initiated to parallel programming. No a priori knowledge on the part of the reader is assumed other than a basic understanding of serial computer programming using a language such as Fortran or C. Anyone may freely access and/or copy this material. The author only requests an appropriate acknowledgement in the event of public reproduction or distribution.

I: Parallel Computing Basics
II: Parallel Computing Hardware
III: Message Passing Interface

A companion document to the parallel computing tutorial is: Dr. Miller's Beowulf Performance Testing Page.

References:

Gropp, W., Lusk, E. and Skjellum, A., `Using MPI: Portable Programming with the Message-Passing Interface,' Second Edition, Massachusetts Institute of Technology Press, 1999.

Gropp, W., Lusk, E. and Thakur, R., `Using MPI-2: Advanced Features of the Message-Passing Interface,' Massachusetts Institute of Technology Press, 1999.

An additional web based resource is the MPICH (MPI Chameleon) implementation homepage maintained at Argonne National Labs:

http://www/mcs.anl.gov/mpi/mpich