In 1 October 2018 we presented our study on how approximated memories impact the energy efficiency of applications in various domains.
Abstract
Approximate memories can lower energy consumption at expense of incurring errors in some of the read/write operations. While these errors may be tolerated in some cases, in general, parts of the application must be re-executed to achieve usable results when a large number of errors occur. Frequent re-executions may, in turn, attenuate or negate energy benefits obtained from using approximate memories. In this work, we show the energy impact of memory approximations in applications considering different quality requirements. Five out of nine selected applications showed a positive energy-quality tradeoff. For these applications, our results show up to 30% energy savings at a 10-8 error rate, when a 20% degradation in quality is allowed.
Content
Full paper PDF (In Portuguese - soon)
Presentation (PDF - In Portuguese)
Link to conference proceedings (soon)
Repositories
VArchC framework
Refer to the documentation for installation instructions.MIPS CPU Model
Refer to the documentation for more information on CPU models.Test applications (soon)
Refer to the documentation for more information on how to run an application and analyze the results.
Cite us
@INPROCEEDINGS{VArchC-WSCAD2018,
author={I. B. Felzmann and J. Fabrício Filho and R. Azevedo and L. F. Wanner},
booktitle={Anais do Simpósio em Sistemas Computacionais de Alto Desempenho (WSCAD)},
title={Impacto de memórias aproximadas na eficiência energética},
year={2018},
}