Concepedia

Performance Comparison of Graphics Processors to Reconfigurable Logic: A Case Study

Ben Cope, Peter Y. K. Cheung, Wayne Luk, Lee Howes

IEEE Transactions on Computers · 2010 · 95 citations · 23 references

TL;DR

The study examines how analysis trends apply to newer and future technologies. The authors define a systematic approach to compare GPUs and reconfigurable logic based on three throughput drivers. They apply this approach to five algorithms differing in arithmetic complexity, memory access, and data dependence, using an NVIDIA GeForce 7900 GTX GPU and a Xilinx Virtex‑4 FPGA. The study finds that both the GPU and FPGA achieve two‑order‑of‑magnitude speedups over a general‑purpose processor for arithmetic‑intensive algorithms, that the FPGA is superior for algorithms with many regular memory accesses while the GPU is better for variable data reuse, and that for data‑dependent algorithms a customized FPGA datapath can outperform the GPU by up to eight times.

Abstract

A systematic approach to the comparison of the graphics processor (GPU) and reconfigurable logic is defined in terms of three throughput drivers. The approach is applied to five case study algorithms, characterized by their arithmetic complexity, memory access requirements, and data dependence, and two target devices: the nVidia GeForce 7900 GTX GPU and a Xilinx Virtex-4 field programmable gate array (FPGA). Two orders of magnitude speedup, over a general-purpose processor, is observed for each device for arithmetic intensive algorithms. An FPGA is superior, over a GPU, for algorithms requiring large numbers of regular memory accesses, while the GPU is superior for algorithms with variable data reuse. In the presence of data dependence, the implementation of a customized data path in an FPGA exceeds GPU performance by up to eight times. The trends of the analysis to newer and future technologies are analyzed.