Solution of the three-dimensional finite difference equations using parallel reconfigurable hardware accelerators

E. J.C. Stewart*, J. Hu, S. F. Quigley, A. H.C. Chan

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper describes a study of the use of FPGA-based reconfigurable hardware, in the form of a coprocessor within a standard PC, to accelerate a solution of the 3-D finite difference formulation of the Laplace equation. The domain decomposition and solution approach are formulated so as to make maximum use of the parallelism within the FPGA, and to minimize limitations imposed by I/O bandwidth between the FPGA and the computer memory. The resulting system provides a red-black SOR solver that can achieve a speed-up of approximately 50 compared to a fast PC.

Original languageEnglish
Title of host publicationProceedings of the 5th International Conference on Engineering Computational Technology
Publication statusPublished - 2006
Event5th International Conference on Engineering Computational Technology, ECT 2006 - Las Palmas de Gran Canaria, Spain
Duration: 12 Sept 200615 Sept 2006

Publication series

NameProceedings of the 5th International Conference on Engineering Computational Technology

Conference

Conference5th International Conference on Engineering Computational Technology, ECT 2006
Country/TerritorySpain
CityLas Palmas de Gran Canaria
Period12/09/0615/09/06

Keywords

  • Domain decomposition
  • Finite difference method
  • FPGAs
  • Hardware acceleration
  • Reconfigurable computing

ASJC Scopus subject areas

  • General Computer Science

Fingerprint

Dive into the research topics of 'Solution of the three-dimensional finite difference equations using parallel reconfigurable hardware accelerators'. Together they form a unique fingerprint.

Cite this