2005 Ohio Student Research Forum

Abstract

Scalability and Efficiency Analysis of a Parallel Algorithm for Simulating Steady State Heat Flow Through a Metal Sheet
Jamon Lewis
Central State University, Department of Computer Science
Mentor: Robert L. Marcus

This research project is to analyzes the scalability and efficiency of a parallel algorithm to solve the steady-state heat flow problem using various grid sizes and various initial conditions. The program simulates heat of 100 o C applied to three edges of a rectangular metal sheet and an ice bath of 00 C applied to the 4th side, and then simulates heat of 100o C applied to two edges of a rectangular metal sheet and an ice bath of 0 o C applied to the other two sides. The programs performs these simulations using a row-wise decomposition method with grid sizes of 100x100 and higher for determining speed-up values using various numbers of nodes. The program simulates the sheet metal insulated so that there is no loss of heat from the surface of the sheet.