Comparing the performance of stochastic simulation on GPUs and OpenMP

Abstract

Since stochastic computing performs operations using streams of bits that represent probability values instead of deterministic values, it can tolerate a large number of failures in a noisy system. However, the simulation of a stochastic implementation is extremely time-consuming. In this paper, we investigate two approaches to speed up the stochastic simulation: a GPU-based simulation and an OpenMP-based simulation. To compare these two approaches, we start with several basic stochastic computing elements (SCEs) and then use the stochastic implementation of a frame difference-based image segmentation algorithm as case study to conduct extensive experiments. Measured results show that the GPU-based simulation with 448 processing elements can achieve up to 119x performance speedup compared to the single-threaded CPU simulation and 17x performance speedup over the OpenMP-based simulation with eight processor cores. In addition, we present several performance optimisations for the GPU-based simulation which significantly benefit the performance of stochastic simulation. Minneapolis. He has worked as a development engineer at Tandem Computers Inc. and has chaired and served on the programme committees of numerous conferences. He was elected a Fellow of the IEEE and the AAAS. His main research interests include computer architecture, computer systems performance analysis, and high-performance storage systems, with a particular interest in the interaction of computer architecture with software, compilers, and circuits. This paper is a revised and expanded version of a paper entitled 'A GPU-based simulation for stochastic computing' presented at the