Abstract: Programming on distributed memory parallel machines involves tremendous efforts on the part of programmers. Without tools support, parallel programming is labor intensive, error-prone, and tedious. The finite element method (FEM) has been widely used for the structural modeling of physical systems. Many research efforts have already provided solid algorithms heading to the sections of FEG partitioning, load balancing, and MPI code packing in various domains. Unfortunately most of them only offer fragmental assistance. The cost of parallel programming is still very high and not cost effective. This paper presents a parallel program generation environment (PPGE) for solving partial differential equations (PDEs) on distributed memory computing environments, especially for clusters of workstations and parallel machines with wormhole-routed interconnection networks. PPGE is aimed to provide an integrated visual parallel programming environment for those users who want to use parallel computers to solve PDEs but do not know how to write the corresponding parallel programs. PPGE has been applied to many real cases of parallel finite element modeling problems. The experimental results show the practicability and usefulness of our PPGE.
Chih-Hung Chang , 2William C. Chu , Chih-Wei Lu , Yeh-Ching Chung and Dong-Ling Yang , 2004. A Parallel Program Generation Environment for Solving PDEs on Distributed Memory Computing Environments . Asian Journal of Information Technology, 3: 1025-1038.