Chinese Journal of Computers   Full Text
Title	A Power Optimization Approach to Real-Time Operating Systems Based on Discrete Hopfield Neural Networks
Authors	GUO Bing1) SHEN Yan2) WANG Dian-Hui3) LI Zhi-Shu1) CHEN Xiang-Dong4)
Address	1)(School of Computer Science & Engineering, Sichuan University, Chengdu 610065) 2)(School of Mechatronics Engineering, University of Electronic Science & Technology of China, Chengdu 610054) 3)(Department of Computer Science & Engineering, La Trobe University, Melbourne, VIC 3086 Australia) 4)(School of Information Science & Technology, Southwest Jiaotong University, Chengdu 610031)
Year	2007
Issue	No.9(1573—1579)
Abstract &   Background	Abstract The RTOS(Real-Time Operating System) is a critical component in the SoC(System-on-a-Chip), which consumes the 30~40% of total system energy in average. Power optimization based on hardware-software partitioning of a RTOS (RTOS-Power partitioning) can significantly reduce the energy consumption of a SoC. This paper presents a new model for RTOS-Power partitioning, which helps in understanding the essence of the RTOS-Power partitioning techniques. A discrete Hopfield neural network approach for implementing the RTOS-Power partitioning is proposed, where a novel neuron expression, energy function, operating equation and coefficients of the neural network are redefined. Simulations are carried out with comparison to generic algorithm and ant algorithm. Experimental results demonstrate that the proposed method can achieve higher energy savings up to 60% at relatively low costs of less than 4K PLBs while increasing the performance compared to the SoC-RTOS realized purely in software. keywords Hopfield neural network; power optimization; RTOS; hardware-software partitioning; SoC background This research is partly supported by the National Natural Science Foundation of China (No.60572026), and undertaken in cooperation with the computation intelligence Laboratory of Department of Computer Science, La Trobe University, Australia. Along with the development of global energy crisis, power issues of software in the vast number of embedded systems have been increasingly concerned. The research group has done much research work in RTOS, software/hardware automatic partitioning algorithms and software energy consumption models. However, RTOS plays an important role in the running of embedded software. Many existed researches mainly concern the dynamic behavior of RTOS, such as power-aware scheduling algorithm. This project mainly focuses on the energy optimization of RTOS(Real-time Operating System) in SoC(System-on-a-Chip) based on software/hardware automatic partitioning method. The authors attempt to employ Hopfield neural network to resolve this combined optimization problem, and achieve some initial results. As a system-level power optimization approach, it can yield high energy savings compared to other techniques, such as Genetic algorithm and Ant algorithm. Simulations experimental results demonstrate that the proposed method can acquire higher energy savings up to 60% at relatively low costs of less than 4K PLBs while increasing the performance compared to the SoC-RTOS realized purely in software.