GPU Versus FPGA for High Productivity Computing
Electrical and Electronic Engineering, Imperial College London, London, UK
International Conference on Field Programmable Logic and Applications (FPL), 2010
@inproceedings{jones2010gpu,
title={GPU versus FPGA for high productivity computing},
author={Jones, D.H. and Powell, A. and Bouganis, C.S. and Cheung, P.Y.K.},
booktitle={2010 International Conference on Field Programmable Logic and Applications},
pages={119–124},
year={2010},
organization={IEEE}
}
Heterogeneous or co-processor architectures are becoming an important component of high productivity computing systems (HPCS). In this work the performance of a GPU based HPCS is compared with the performance of a commercially available FPGA based HPC. Contrary to previous approaches that focussed on specific examples, a broader analysis is performed by considering processes at an architectural level. A set of benchmarks is employed that use different process architectures in order to exploit the benefits of each technology. These include the asynchronous pipelines common to "map" tasks, a partially synchronous tree common to "reduce" tasks and a fully synchronous, fully connected mesh. We show that the GPU is more productive than the FPGA architecture for most of the benchmarks and conclude that FPGA-based HPCS is being marginalised by GPUs.
September 2, 2011 by hgpu
