Scalable Implementation of the Discrete Element Method on a Reconfigurable Computing Platform

Abstract

The Discrete Element Method (DEM) is a numerical method for the simulation of the behaviour of media consisting of discrete particles, and is important for the optimisation of production processes in the food processing and pharmaceutical industries. The DEM is computationally very expensive, but can benefit from the properties of reconfigurable computing. This paper presents the design of an optimised dedicated hardware architecture for the DEM implemented on a Reconfigurable Computing platform based on a Field Programmable Gate Array. A single FPGA version shows a 30-fold speed-up compared to an optimised software version running on a fast PC. The design of a multi-FPGA board solution is also presented; this provides approximately 30 times more speed up for each FPGA used.