Table of Contents Author Guidelines Submit a Manuscript
Scientific Programming
Volume 6, Issue 1, Pages 95-113
http://dx.doi.org/10.1155/1997/174748

Automatic Data Distribution for Composite Grid Applications

Lorie M. Liebrock1 and Ken Kennedy2

1Liebrock-Hicks Research, 2315 County Road, Calumet, MI 49913, USA
2Department of Computer Science, Rice University, Houston, TX 77251, USA

Received 26 December 1994; Accepted 26 February 1996

Copyright © 1997 Hindawi Publishing Corporation. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Problem topology is the key to efficient parallelization support for partially regular applications. Specifically, problem topology provides the information necessary for automatic data distribution and regular application optimization of a large class of partially regular applications. Problem topology is the connectivity of the problem. This research focuses on composite grid applications and strives to take advantage of their partial regularity in the parallelization and compilation process. Composite grid problems arise in important application areas, e.g., reactor and aerodynamic simulation. Related physical phenomena are inherently parallel and their simulations are computationally intensive. We present algorithms that automatically determine data distributions for composite grid problems. Our algorithm's alignment and distribution specifications may be used as input to a High Performance Fortran program to apply the mapping for execution of the simulation code. These algorithms eliminate the need for user-specified data distribution for this large class of complex topology problems. We test the algorithms using a number of topological descriptions from aerodynamic and water-cooled nuclear reactor simulations. Speedup-bound predictions with and without communication, based on the automatically generated distributions, indicate that significant speedups are possible using these algorithms.