Comparing performance and energy efficiency of FPGAs and GPUs for high productivity computing

Brahim Betkaoui, David B. Thomas, Wayne Luk

2010 · 87 citations · 22 references

Concepts

TL;DR

Interest in heterogeneous systems combining FPGA and GPU accelerators has grown due to their performance gains, but their high design effort reduces programmer productivity, prompting high‑productivity approaches such as CUDA for GPUs. The study compares the performance and energy efficiency of high‑productivity computing systems based on FPGA and GPU technologies, specifically evaluating the High‑Productivity Reconfigurable Computer (HPRC) FPGA programming approach. The authors benchmarked an FPGA‑based Hybrid‑Core system and a GPU system on a suite of established benchmarks with varying memory access patterns to assess performance and energy efficiency. Results indicate GPUs excel at streaming workloads, whereas high‑productivity reconfigurable computing systems outperform GPUs on applications with poor locality and low memory bandwidth.

Abstract

This paper provides the first comparison of performance and energy efficiency of high productivity computing systems based on FPGA (Field-Programmable Gate Array) and GPU (Graphics Processing Unit) technologies. The search for higher performance compute solutions has recently led to great interest in heterogeneous systems containing FPGA and GPU accelerators. While these accelerators can provide significant performance improvements, they can also require much more design effort than a pure software solution, reducing programmer productivity. The CUDA system has provided a high productivity approach for programming GPUs. This paper evaluates the High-Productivity Reconfigurable Computer (HPRC) approach to FPGA programming, where a commodity CPU instruction set architecture is augmented with instructions which execute on a specialised FPGA co-processor, allowing the CPU and FPGA to co-operate closely while providing a programming model similar to that of traditional software. To compare the GPU and FPGA approaches, we select a set of established benchmarks with different memory access characteristics, and compare their performance and energy efficiency on an FPGA-based Hybrid-Core system with a GPU-based system. Our results show that while GPUs excel at streaming applications, high-productivity reconfigurable computing systems outperform GPUs in applications with poor locality characteristics and low memory bandwidth requirements.

References

22