Virtual data space - load balancing for irregular applications

Load balancing is a key issue in the development of parallel algorithms wit h irregular structures. Existing load balancing systems each support only o ne specific programming paradigm and thus are of limited use. The system VD S presented here allows concurrent use of various paradigms such as fork-jo in, weighted tasks, and static dags (directed acyclic graphs that are known in advance). The system provides visual performance evaluation tools to fa cilitate the efficient application of the system. VDS supports various comm unication interfaces including PVM and MPI. Thus, VDS-applications can be r un on architectures ranging from workstation clusters to massively parallel systems. (C) 2000 Elsevier Science B.V. All rights reserved.