Skip navigation, to content.

2006 Rochester Computational Science and Education Conference

Domain decomposed parallel heat distribution problem in two dimensions

Author: Y. Kortsarts (Widener University, Computer Science Dept.)

Abstract

Finite difference method is widely used in scientific and engineering computing to solve varieties of problems. We have used the finite difference method to solve the heat equation in an isotropic two dimensional material. A simple parallel algorithm using domain decomposition technique was implemented and tested in a cluster of computers. We present the benchmarking results of these programs.

Performance study of domain decomposed parallel matrix-vector multiplication programs

Author: Y. Kortsarts (Widener University, Computer Science Dept.)

Abstract

Dense matrix-vector multiplication algorithm has many scientific and engineering applications, which includes image processing and neural networks. Fast execution of matrix vector multiplication algorithm using parallel computers is therefore needed. In this paper we report the results of a performance study of three domain-decomposed parallel programs. The programs were implemented and tested in a cluster of computers.