Power Reduction of Automatic Heterogeneous Device Offloading

TL;DR

This paper presents an environment-adaptive software approach that automatically offloads tasks to heterogeneous hardware like GPU and FPGA, optimizing for low power consumption and verifying performance improvements through power utilization comparisons.

Contribution

It introduces a novel environment-adaptive system for automatic hardware offloading that simplifies heterogeneous computing and reduces power consumption.

Findings

Automatic offloading improves power efficiency.

Power utilization decreases with environment-adaptive offloading.

Performance gains confirmed through Watt*seconds comparison.

Abstract

In recent years, utilization of heterogeneous hardware other than small core CPU such as GPU, FPGA or many core CPU is increasing. However, when using heterogeneous hardware, barriers of technical skills such as CUDA are high. Based on that, I have proposed environment-adaptive software that enables automatic conversion, configuration, and high performance and low power operation of once written code, according to the hardware to be placed. I also have verified performance improvement of automatic GPU and FPGA offloading so far. In this paper, I verify low power operation with environment adaptation by confirming power utilization after automatic offloading. I compare Watt*seconds of existing applications after automatic offloading with the case of CPU only processing.