Numerical Simulation of the Frank-Kamenetskii PDE: GPU vs. CPU Computing
Faculty of Science, University of the Witwatersrand, School of Computational and Applied Mathematics, Centre for Differential Equations, Continuum Mechanics and Applications, South Africa
Chapter in book "MATLAB – A Fundamental Tool for Scientific Computing and Engineering Applications – Volume 3", Edited by Vasilios N. Katsikis, 2012
DOI:10.5772/46463
@article{harley2012numerical,
title={Numerical Simulation of the Frank-Kamenetskii PDE: GPU vs. CPU Computing},
author={Harley, C.},
year={2012}
}
The efficient solution of the Frank-Kamenetskii partial differential equation through the implementation of parallelized numerical algorithms or GPUs (Graphics Processing Units) in MATLAB is a natural progression of the work which has been conducted in an area of practical import. There is an on-going interest in the mathematics describing thermal explosions due to the significance of the applications of such models – one example is the chemical processes which occur in grain silos. Solutions which pertain to the different geometries of such a physical process have different physical interpretations, however in this chapter we will consider the Frank-Kamenetskii partial differential equation within the context of the mathematical theory of combustion which according to Frank-Kamenetskii [16] deals with the combined systems of equations of chemical kinetics and of heat transfer and diffusion. A physical explanation of such a system is often a gas confined within a vessel which then reacts chemically, heating up until it either attains a steady state or explodes. The focus of this chapter is to investigate the performance of the parallelization power of the GPU vs. the computing power of the CPU within the context of the solution of the Frank-Kamenetskii partial differential equation. GPU computing is the use of a GPU as a co-processor to accelerate CPUs (Central Processing Units) for general purpose scientific and engineering computing. The GPU accelerates applications running on the CPU by offloading some of the compute-intensive and time consuming portions of the code. The rest of the application still runs on the CPU. The reason why the application is seen to run faster is because it is using the extreme parallel processing power of the GPU to boost performance. A CPU consists of 4 to 8 CPU cores while the GPU consists of 100s of smaller cores. Together they operate to crunch through the data in the application and as such it is this massive parallel architecture which gives the GPU its high compute performance.
November 1, 2012 by hgpu
