Author(s)

S. U. Ewedafe & R. H. Shariffudin

Abstract

In this paper, a method based on applying Successive Over-Relaxation (SOR) to find the solution of 2-Dimensional Bio-Heat Transfer Problem (2DBHTP) on a distributed computing environment using Input File Affinity Measure (Iaff) with Message Passing Interface (MPI) is introduced. A finite difference discretization is developed to solve the 2DBHTP since 2DBHTP requires the evaluation of temporal and spatial distributions of temperature. The platform gives us better predictions of the effects of thermal physical properties of temperature distribution. This work proposes a parallel overhead with overlapping communication in its implementation using MPI. Based on the implementation of the MPI distributed computing environment, we present a performance improvement strategy running on the master-slave platform. Parallel comparisons of the method were experimentally evaluated, and parallel results show speedup and efficiency on a various number of mesh sizes. The method used combines elements of numerical stability and parallel algorithm design. Keywords: 2DBHTP, SOR, Iaff, MPI, speedup, discretization, finite difference, spatial temperature, parallel overhead, performance.

Keywords

Keywords: 2DBHTP, SOR, Iaff, MPI, speedup, discretization, finite difference, spatial temperature, parallel overhead, performance.