Paper Info
Title
Energy-Aware Automatic Tuning on Many-Core Platform via Differential Evolution
Abstract
The need for energy-aware computing has become increasingly demanding. However, high-performance computing is always users' pursuit. The key is how to maintain high performance meanwhile consuming an acceptable amount of energy. One feasible method is the energy-aware automatic tuning. In this paper, we implemented and compared several Differential Evolution (DE) based auto-tuning algorithms on an innovative many-core platform, in an effort to improve the energy consumption and energy delay product (EDP) of the entire platform. Our results show that Adaptive Differential Evolution algorithm is able to achieve better energy consumption as well as EDP than other DE algorithms tested. It also converges faster. We believe DE algorithm is an effective method that can be applied towards energy-efficient computing on manycore platforms.
Year
DOI
Venue
2016
10.1109/ICPPW.2016.46
2016 45th International Conference on Parallel Processing Workshops (ICPPW)
Keywords
Field
DocType
Differential Evolution,Dynamic Voltage and Frequency Scaling,Energy-Aware Computing,Many-core Processors
Computer science,Effective method,Parallel computing,Differential evolution,Automatic tuning,Energy consumption,Differential evolution algorithm,Distributed computing
Conference
ISSN
ISBN
Citations 
1530-2016
978-1-5090-2826-9
0
PageRank 
References 
Authors
0.34
8
3
Name
Order
Citations
PageRank
Yijun Jiang101.69
Xuan Qi262.46
Chen Liu38216.75