Parallel Computing with Low-Cost FPGAs: A Framework for COPACOBANA

Tim Güneysu, Chris­tof Paar, Jan Pelzl, G. Pfeiffer, M. Schimmler, C. Schleiffer

Parallel Computing: Architectures, Algorithms and Applications, ParCo 2007, Forschungszentrum Jülich and RWTH Aachen University, Germany, 4-7 September 2007


Abstract

In many disciplines such as applied sciences or computer science, computationally challenging problems demand for extraordinary computing power, mostly provided by super computers or clusters of conventional desktop CPUs. During the last decades, several ?avors of super computers have evolved, most of which are suitable for a speci?c type of problem. In general, dedicated clusters and super computers suffer from their extremely high cost per computation and are, due to the lack of cheap alternatives, currently the only possible solution to computational hard problems. More recently, emerging low-cost FPGAs tend to be a very cost-effective alternative to conventional CPUs for solving at least some of the computational hard problems such as those appearing in cryptanalysis and bio-informatics

[BibTeX] [pdf]

Tags: