Title: A framework for comparing high performance computing technologies

Authors: Randall E. Duran; Ding Chen; Rishi Saraswat; Aaron Hallmark

Addresses: School of Information Systems, Singapore Management University, 80 Stamford Road, 178902, Singapore ' Catena Technologies Pte Ltd., 30 Robinson Road, Robinson Towers #11-04, 048546, Singapore ' Catena Technologies Pte Ltd., 30 Robinson Road, Robinson Towers #11-04, 048546, Singapore ' Catena Technologies Pte Ltd., 30 Robinson Road, Robinson Towers #11-04, 048546, Singapore

Abstract: Over the past decade many different technologies have become available for accelerating applications and implementing high performance computing solutions. However, the abundance of technologies and novel architectures available creates a new challenge of effectively selecting the optimal one to achieve specific project goals. This research examines the challenges of quantifying technology characteristics and application requirements, and introduces a framework for comparing and assessing the most appropriate technology. To illustrate how the framework can be applied, two use case examples related to financial services are presented that compare the suitability of multi-core CPU, GPU, and FPGA technologies.

Keywords: high performance computing; HPC; technology selection; field programmable gate arrays; FPGA; graphics processing unit; GPU; multi-core processors; technology comparison framework.

DOI: 10.1504/IJCSE.2014.058704

International Journal of Computational Science and Engineering, 2014 Vol.9 No.1/2, pp.119 - 129

Received: 20 Jan 2012
Accepted: 17 Apr 2012
Published online: 24 May 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article