High-Performance Hardware Accelerators for Solving Ordinary Differential Equations

Conference paper (2017)

Authors

Ioannis Stamoulias National and Capodistrian University of Athens
M. Möller Numerical Analysis - EEMCS
Rene Miedema Student
C. Strydis Erasmus MC
Christoforos Kachris National Technical University of Athens
D. Soudris National Technical University of Athens
Research Group
Numerical Analysis (EEMCS) (TU Delft)
More Info
expand_more

Published Date

2017

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Electrical Engineering, Mathematics and Computer Science

Department

Delft Institute of Applied Mathematics

Research Group

Numerical Analysis

Abstract

Ordinary Differential Equations (ODEs) are widely used in many high-performance computing applications. However, contemporary processors generally provide limited throughput for these kinds of calculations. A high-performance hardware accelerator has been developed for speeding-up the solution of ODEs. The hardware accelerator has been developed both for single and double floating-point precision types and a design-space exploration has been performed in terms of performance and hardware resources. The hardware accelerator has been mapped to an FPGA board and connected through PCIe to a typical processor. The performance evaluation shows that the proposed scheme can achieve up to 14x speedup compared to a reference, single-core CPU solution.