Simulation of a Model for Refactoring Approach for Parallelism Using Parallel Computing Tool Box

Abstract

Refactoring is the process of retaining the behavior of a program by making changes to the structure of a program. Initially refactoring is used only for sequential programs, but due to highly configurated architectural availability, it also aids parallel programmers in implementing their parallel applications. Refactoring provides many advantages to parallel programmers, in identifying independent modules, in refining process of programs, it also helps in separating concerns between application and system programmers, and it reduces the time for deployment. All mentioned advantages benefit the programmer in writing parallel programs. The approach for refactoring using multi core system is already developed. Hence all these advantages made us to thought of a system to develop refactoring approach for parallelism which uses heterogeneous parallel architectures which uses combination of both Graphic Processing Unit (GPU) and Central Processing Unit (CPU). A Tool in MATLAB, Parallel Computing Toolbox can be used to execute programs on multiple processing elements simultaneously with local workers available in the toolbox, which takes benefit of GPUs. This tool box uses complete processing speed of multi core system to execute applications on local workers without changing the code. Our suggested model can be simulated by using Parallel Computing Toolbox.