Performance-Aware Power Management in Embedded Controllers with Multiple-Voltage Processors

Name of Corresponding Author: Feng Xia Complete Postal Address: College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China E-mail: f.xia@ieee.org Alternate E-mail: f.xia@acm.org

Abstract

The goal of this work is to minimize the energy dissipation of embedded controllers without jeopardizing the quality of control (QoC). Taking advantage of the dynamic voltage scaling (DVS) technology, this paper develops a performance-aware power management scheme for embedded controllers with processors that allow multiple voltage levels. The periods of control tasks are adapted online with respect to the current QoC, thus facilitating additional energy reduction over standard DVS. To avoid the waste of CPU resources as a result of the discrete voltage levels, a resource reclaiming mechanism is employed to maximize the CPU utilization and also to improve the QoC. Simulations are conducted to evaluate the performance of the proposed scheme. Compared with the optimal standard DVS scheme, the proposed scheme is shown to be able to save remarkably more energy while maintaining comparable QoC.

Keyword: Energy management, embedded systems, application adaptation, real-time control

Introduction Applications of embedded processors in various engineering systems have been expanding rapidly in recent years. With continuous miniaturization of physical size, an increasing number of embedded processors are battery powered. For these energy-limited systems, power management is important because low power dissipation not only prolongs the battery’s lifetime but also increases the system reliability. Even if the energy constraint does not exist, energy should also be conserved to reduce the operational cost and environmental effect.
With the advent of mobile computing and wireless networking techniques, the use of battery- powered embedded processors in real-time control systems is rapidly growing. A typical example is mobile robots. Consequently, control engineers are confronted with a ‘new’ type of resource management problem, i.e., power management. In general, however, low energy consumption and Performance-Aware Power Management in Embedded Controllers with Multiple-Voltage Processors 1,4Feng Xia, 2Liping Liu, 3Longhua Ma, 3Youxian Sun and 1Jinxiang Dong 1College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China 2School of Electrical Engineering and Automation, Tianjin University, Tianjin 300072, China 3State Key Lab of Industrial Control Technology, Zhejiang University, Hangzhou 310027, China 4Faculty of Information Technology, Queensland University of Technology, Brisbane QLD 4001, Australia

2 high-quality system performance are conflicting with each other (Pillai and Shin, 2001). Therefore, in embedded controllers the energy consumption must be minimized in a way that the Quality-of- Control (QoC) of the system is not jeopardized. Significant effort has been made on power and energy management in general-purpose embedded systems. The majority of the achieved results take advantage of the dynamic voltage scaling (DVS) technology (Aydin et al., 2004; Gaujal and Navet, 2007). By adjusting the operating voltage and frequency of the processors dynamically, DVS has proved effective in energy conservation since the energy dissipation is approximately proportional to the square of the voltage (i.e., E∝V2). Regardless, limited work has been done in power management for real-time control applications. As one of the first work in this direction, Lee and Kim (2005) formulated the power management in multitasking control systems as an optimization problem, and proposed a static solution and a dynamic solution for the problem. Jin et al. (2007) presented a feedback fuzzy-DVS scheduling architecture integrating feedback control and fuzzy DVS for real-time control tasks. Recently, we have also developed several DVS algorithms for real-time control systems (Xia and Sun, 2006; Xia et al., 2008).
A common assumption of the above-mentioned approaches is that the supply voltage of the processor could be varied continuously. However, this is not the case for modern processors: only a limited number of voltage levels are practically available. With these multiple-voltage processors, applying DVS algorithms that assume continuous voltage levels to embedded control systems cannot realize the full potential of energy reduction (Hua and Qu, 2005). To guarantee the system schedulability, a waste of computing resources may potentially result from the quantization of the voltage levels. Recently, Marinoni and Buttazzo (2007) proposed a method th

…(Full text truncated)…