智能移动嵌入式系统的协同节能关键技术研究

项目名称： 智能移动嵌入式系统的协同节能关键技术研究

项目编号： No.61272104

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 郭兵

作者单位： 四川大学

项目金额： 85万元

中文摘要： 在目前全球倡导"低碳节能"的背景下，面对移动互联网快速发展的需要，智能移动嵌入式系统（简称智能嵌入式系统，如智能手机）的能耗问题始终是困扰智能嵌入式系统高性能应用与多样化服务的重要瓶颈。为提高智能嵌入式系统的续航能力与可用性，本研究从硬件单元功耗、软件任务运行效率及系统资源配置策略三个方面探讨智能嵌入式系统的协同节能技术，首先研究新型高能效的可定制异构多核体系结构，包括可定制核和可编程模块的高能效异构多核结构和能量感知的编程模式等软/硬件两个层面的静态低功耗设计。然后，基于任务的动态性和差异性，设计一种基于电池电量反馈的自适应闭环控制任务调度模型，实现能量可感知的任务运行模式，并根据任务运行的"局部化"原理，形成一种软/硬件资源的自适应节能配置策略，三个方面节能技术的交互与协同，以获得能量资源全局最优的应用效率。最后，进一步探讨能耗与速度等指标关系，并集成研究成果开展实例应用与验证工作。

中文关键词： 嵌入式系统能耗；协同节能；能耗分析；能耗优化；能耗评价

英文摘要： Nowadays as low carbon economy is advocated worldwide, aimed for the requirements of mobile INTERNET rapid development, the energy consumption is alaways a crucial issue to seriously restrain the high-performance applications and diverse service of smart, portable embedded systems (namely smart embedded systems, e.g., smart phone). To significantly improve the standby time and service availability of smart embedded systems, this research explores the collaboration energy-saving technologies from three aspects,i.e.,the power of hardware units, runtime efficiency of software tasks and configuration policies in hardware/software resources. Firstly, this research analyzes the architecture of new high energy-efficiency heterogeneous multi-cores, including the two-layer static low-power design of hardware and software, i.e., the high power-efficiency heterogeneous multi-core structure with some customized cores and programmable modules, and energy-aware software programming mode. Then, a self-adapted loop-back control task scheduling model based on the battery power feedback is design to implement the energy-aware task execution mode, and a hardware/software resource self-adapted energy-saving configuration policies based on the localization principles of task running is constructed. These three aspects of energy-savi

英文关键词： embedded system energy consumption；collaboration energy saving；energy consumption analysis；energy consumption optimization；energy consumption evaluation