源码未知类软件能耗评估技术研究
|
摘要
目前，能源危机和环境问题成为全世界普遍关注的问题，各行各业都致力于节能减排研究。在计算机领域，低能耗设计起源于计算机硬件，不断涌现出各种节能硬件和绿色计算机产品。随着硬件低能耗设计技术的日趋成熟，软件能耗逐渐成为影响计算机系统能耗的重要因素。国内外学者提出很多分析和优化软件能耗的方法，但所有研究都是以软件源代码已知为基础的，关于源码未知类软件能耗的研究仍属空白。而日常所用的软件大多数都是源代码未知的，包括各种商业软件和大量网络下载的免费软件。源码未知类软件通常没有规范化的软件开发和测试过程，无法保证软件设计过程遵循了低能耗软件设计原则，目前也没有有效的理论和方法对其能耗问题进行分析和评估。这些问题导致源码未知类软件在一定程度上产生无效能耗，另外此类软件使用范围广、数量大，造成的能源浪费是不容忽视的。
针对这一问题，本文对源码未知类软件能耗问题进行分析，试图找到源码未知类软件能耗评估方法。操作系统规模大、结构复杂并且与计算机硬件直接接触，因此本文分别对源码未知类应用软件和源码未知类操作系统能耗评估问题进行研究。源码未知类软件具有黑盒特性，因此本文研究以实际测量为基础。搭建测量平台并编写相应辅助软件，研究过程中能耗的测量和分析由测量平台辅助完成。最终实现对源码未知类软件建立通用能耗评估模型，对源码未知类操作系统在理论层次建立能耗对比评估模型。
深入分析国内外主要的软件能耗分析技术和优化方法，找到源码未知类应用软件能耗分析的切入点。分析结果显示：无论采用哪种技术，低能耗优化最终都体现在编译器、数据结构和算法的选择上。针对源码未知类软件能耗问题，提出根源因素和典型外部特征两个新的概念。结合理论分析和实验验证，论证上述三个因素为影响软件能耗的根源因素，并找到与根源因素对应的典型外部特征。以播放器软件为例进行能耗测量，以测量数据为基础对典型外部特征与软件能耗之间的关系进行定性、定量分析，实验结果表明功能相同的不同软件具有不同的能耗特性。
依据综合评估问题的不明确性，将模糊理论应用于软件能耗评估领域。基于源码未知类软件能耗特性，以模糊理论为基础建立源码未知类应用软件能耗评估模型。利用模糊综合评判、模糊模式识别以及模糊决策等理论解决模型建立过程中的关键性问题。模型在对软件性质进行综合评估的基础上，对软件能耗进行合理评估。该模型为通用模型，如果具备了完善的能耗评估准则库，则可以对任意一款源码未知类软件能耗进行评估。在本文测试数据基础上，利用模型对播放器软件进行软件能耗评估。
对源码未知类操作系统能耗评估问题进行初步探索，在理论层次提出操作系统能耗对比评估模型，为后续研究提供基础。以Windows XP、Windows Vista操作系统为例，对计算机启动过程、待机状态以及基本文件操作进行能耗测试，验证操作系统间能耗差异的存在性。论证操作系统能耗评估的重要性，深入分析操作系统能耗评估问题的复杂性，针对操作系统之间的不同主要体现在核心技术和功能两个方面，基于操作系统功能和操作系统低功耗技术提出源码未知类操作系统的能耗对比评估模型。模型包括功能能耗测试模块、低能耗技术评估测试模块。以动态功耗管理技术为例探索低功耗技术效果评估方法，提出一种新的动态调整超时策略，并以此为基础建立动态功耗管理评估模型。
To confront of energy crisis and environment problems, all walks of life committhemselves to energy saving and emission reduction. In computer science,technology of low power design for hardware is mature. According the appearanceof “green” hardware, software cost is the major component for power consumptionof computer system. There are lots of researches about software cost. Most of themanalyze software cost based on open source. There is no research about source codesealed software. But most of daily software is source code sealed, includingcommercial software and free software downloaded from Internet. There is noprocess of software testing and no theory or method to evaluate power consumptionfor this kind of software. No one knows whether the principle of low power design isused or not during the process of software design. This kind of software will produceinvalid power consumption and waste energy.
To handle above problem, technology of power evaluating for source codesealed software is studied in this paper. Application software and operating softwareare researched respectively. Source code sealed software is black box, so research inthis paper is based on actual testing. Testing bed is built and auxiliary software iscoded. They are responsible for testing and analyzing power consumption.
For application software, study traditional analyzing and optimizing methodsfor software cost, and find out the entry point of research about source code sealed software. They are compiler, data structure and algorithm. Analyze in theory and testcorresponding program in testing bed to demonstrate they are root factors. At thesame time, summarize the classical outer characters corresponding to root factorsfrom experimental results. Take video player as an example to analyze therelationship between characters and software cost qualitatively and quantitatively.Results show that different software has different power character.
The problem of comprehensive evaluating has the character of fuzzy. So powerevaluating model for source code sealed software is based on fuzzy theory. There areevaluating for software characters, evaluating for power consumption and evaluatingdatabase. Fuzzy comprehensive evaluation, fuzzy pattern recognition and fuzzydecision making is used to build evaluating model. Evaluate power consumption forvideo player using model proposed in this paper.
Preliminary exploration of power evaluating for source code sealed operatingsystem is studied in this paper. A power evaluating model is proposed in theory. Itwill be foundation of future research. Test power consumption of process of booting,monitor and operating for file management, and verify the existence of the differencefor power consumption between different operating systems. Introduce theimportance of research of power evaluating for operating system, and analyze itscomplexity. The difference between operating systems reflects on core technologyand functions. Aiming to the two kinds of difference, a power evaluating model isproposed based on function and low power technology of operating system. Themodel is composed of power evaluating of function and power evaluating of lowpower technology. Take dynamic power management as example to study thetechnology of evaluating low-power design methods. Propose a new time-out policyand put it into evaluating model for dynamic power management.
针对这一问题，本文对源码未知类软件能耗问题进行分析，试图找到源码未知类软件能耗评估方法。操作系统规模大、结构复杂并且与计算机硬件直接接触，因此本文分别对源码未知类应用软件和源码未知类操作系统能耗评估问题进行研究。源码未知类软件具有黑盒特性，因此本文研究以实际测量为基础。搭建测量平台并编写相应辅助软件，研究过程中能耗的测量和分析由测量平台辅助完成。最终实现对源码未知类软件建立通用能耗评估模型，对源码未知类操作系统在理论层次建立能耗对比评估模型。
深入分析国内外主要的软件能耗分析技术和优化方法，找到源码未知类应用软件能耗分析的切入点。分析结果显示：无论采用哪种技术，低能耗优化最终都体现在编译器、数据结构和算法的选择上。针对源码未知类软件能耗问题，提出根源因素和典型外部特征两个新的概念。结合理论分析和实验验证，论证上述三个因素为影响软件能耗的根源因素，并找到与根源因素对应的典型外部特征。以播放器软件为例进行能耗测量，以测量数据为基础对典型外部特征与软件能耗之间的关系进行定性、定量分析，实验结果表明功能相同的不同软件具有不同的能耗特性。
依据综合评估问题的不明确性，将模糊理论应用于软件能耗评估领域。基于源码未知类软件能耗特性，以模糊理论为基础建立源码未知类应用软件能耗评估模型。利用模糊综合评判、模糊模式识别以及模糊决策等理论解决模型建立过程中的关键性问题。模型在对软件性质进行综合评估的基础上，对软件能耗进行合理评估。该模型为通用模型，如果具备了完善的能耗评估准则库，则可以对任意一款源码未知类软件能耗进行评估。在本文测试数据基础上，利用模型对播放器软件进行软件能耗评估。
对源码未知类操作系统能耗评估问题进行初步探索，在理论层次提出操作系统能耗对比评估模型，为后续研究提供基础。以Windows XP、Windows Vista操作系统为例，对计算机启动过程、待机状态以及基本文件操作进行能耗测试，验证操作系统间能耗差异的存在性。论证操作系统能耗评估的重要性，深入分析操作系统能耗评估问题的复杂性，针对操作系统之间的不同主要体现在核心技术和功能两个方面，基于操作系统功能和操作系统低功耗技术提出源码未知类操作系统的能耗对比评估模型。模型包括功能能耗测试模块、低能耗技术评估测试模块。以动态功耗管理技术为例探索低功耗技术效果评估方法，提出一种新的动态调整超时策略，并以此为基础建立动态功耗管理评估模型。
To confront of energy crisis and environment problems, all walks of life committhemselves to energy saving and emission reduction. In computer science,technology of low power design for hardware is mature. According the appearanceof “green” hardware, software cost is the major component for power consumptionof computer system. There are lots of researches about software cost. Most of themanalyze software cost based on open source. There is no research about source codesealed software. But most of daily software is source code sealed, includingcommercial software and free software downloaded from Internet. There is noprocess of software testing and no theory or method to evaluate power consumptionfor this kind of software. No one knows whether the principle of low power design isused or not during the process of software design. This kind of software will produceinvalid power consumption and waste energy.
To handle above problem, technology of power evaluating for source codesealed software is studied in this paper. Application software and operating softwareare researched respectively. Source code sealed software is black box, so research inthis paper is based on actual testing. Testing bed is built and auxiliary software iscoded. They are responsible for testing and analyzing power consumption.
For application software, study traditional analyzing and optimizing methodsfor software cost, and find out the entry point of research about source code sealed software. They are compiler, data structure and algorithm. Analyze in theory and testcorresponding program in testing bed to demonstrate they are root factors. At thesame time, summarize the classical outer characters corresponding to root factorsfrom experimental results. Take video player as an example to analyze therelationship between characters and software cost qualitatively and quantitatively.Results show that different software has different power character.
The problem of comprehensive evaluating has the character of fuzzy. So powerevaluating model for source code sealed software is based on fuzzy theory. There areevaluating for software characters, evaluating for power consumption and evaluatingdatabase. Fuzzy comprehensive evaluation, fuzzy pattern recognition and fuzzydecision making is used to build evaluating model. Evaluate power consumption forvideo player using model proposed in this paper.
Preliminary exploration of power evaluating for source code sealed operatingsystem is studied in this paper. A power evaluating model is proposed in theory. Itwill be foundation of future research. Test power consumption of process of booting,monitor and operating for file management, and verify the existence of the differencefor power consumption between different operating systems. Introduce theimportance of research of power evaluating for operating system, and analyze itscomplexity. The difference between operating systems reflects on core technologyand functions. Aiming to the two kinds of difference, a power evaluating model isproposed based on function and low power technology of operating system. Themodel is composed of power evaluating of function and power evaluating of lowpower technology. Take dynamic power management as example to study thetechnology of evaluating low-power design methods. Propose a new time-out policyand put it into evaluating model for dynamic power management.
引文
[1]郭兵，沈艳，邵子力．绿色计算的重定义与若干探讨[J]．计算机学报，2009，32(12)：2311-2319
[2]林闯，田源，姚敏．绿色网络和绿色评价：节能机制、模型和评价[J]．计算机学报，2011，34(4)：594-612
[3]王怀民，史佩昌，丁博，尹刚，史殿习．软件服务的在线演化[J].计算机学报，2011，2(34)：318-327
[4]刘欣洋．复杂网络环境下基于模型驱动的软件进化方法研究[D]：[硕士学位论文]．呼和浩特：内蒙古大学计算机学院，2011
[5]杜思奇.面向普适计算的软件动态更新[D]：[硕士学位论文]．上海：上海交通大学计算机科学与工程系，2011
[6] V. Tiwari, S. Malik, et al. Power analysis of embedded software: a first steptowards software power minimization [J]. IEEE Transactions on Very Large ScaleIntegration Systems (VLSI),1994,2(4):437-445
[7] V. Tiwari, S. Malik, A. Wolfe. Instruction Level power analysis andoptimization of software [J]. Journal of VLSL Signal Processing,1996(13):223-238
[8]陈嘉，董渊，杨阳，戴桂兰，王生原．基于指令聚类与指令调度的嵌入式软件功耗优化研究[J]．小型微型计算机系统，2006，1(27)：175-179
[9] C. Lee, J. K. Lee, T. T. Hwang. Compiler optimization on instruction schedulingfor low power [C]. ISSS2000,2000:20-22
[10]赵荣彩，唐志敏，张兆庆，Guang R. Gao．低功耗多线程编译优化技术[J]．软件学报，2002，13(6)：1123-1129
[11]赵荣彩，唐志敏，张兆庆，Guang R. Gao．软件流水的低功耗编译技术研究[J]．软件学报，2003，14(8)：1357-1363
[12]易会战，杨学军．有效的低功耗编译优化方法：部件使用局部化[J]．软件学报，2004，15(10)：1451-1460
[13] D. Vishal, C. P. Ravikumar. Software power optimization in an embeddedsystem [C]. Proceedings of the14th International Conference on VLSI Design,2000:254-259
[14] Kyu-won choi, Abhijit Chatterjee. Efficient instruction-level optimizationmethodology for low power embedded systems [C]. ACM Proceeding of theInternational Symposium on Synthesis,2001:147-152
[15] Brandolese Carlo. Source level estimation of energy consumption and executiontime of embedded software [C]. Proceedings of the11th EUROMICRO Conferenceon Digital System Design Architectures, Methods and Tools,2008:115-123
[16]廖海艳，范明明．源码级和算法级的功耗测试与优化[J]．单片机与嵌入式系统应用，2010，1：11-14
[17]罗刚，郭兵，沈艳，廖海艳，任磊．源程序级和算法级嵌入式软件功耗特性的分析与优化方法研究[J]．计算机学报，2009，32(9)：1869-1875
[18] T. Mukherjee, Q. H. Tang, C. Ziesman, et al. Software architecture for dynamicthermal management in datacenters [C]. Proceedings of the2nd InternationalConference on Communication Systems Software and Middle-ware,2007:1-11
[19] V. Swaminathan, K. Chakrabarty. Network flow techniques for dynamic voltagescaling in hard real-time systems [J]. IEEE Transactions on Computer-aided Design ofIntergrated Circuits and Systems,2004,23(10):1385-1398
[20]易会战，陈娟，杨学军，刘喆．基于语法树的实时动态电压调节低功耗算法[J]．软件学报，2005，16(10)：1726-1734
[21]吴琦，熊光泽．非平稳自相似业务下自适应动态功耗管理[J]．软件学报，2005，16(8)：1499-1505
[22]吴琦，熊光泽．基于随机决策模型的动态功耗管理策略研究[J]．计算机学报，2007，30(4)：622-628
[23] Q. Wu, G. Z. Xiong. Why simple timeout strategies work perfectly in practice[J]. Lecture Notes in Computer Science,2006(4221):555-564
[24] E. Y. Chung, L. Benini, et al. Dynamic power management for nonstationaryservice requests [J]. IEEE Transactions on Computers,2002,51(11):1345-1360
[25] S. Li, C. Zeng, et al. An adaptive dynamic power management predictive policy[J]. Information and Electronic Engineering,2010,8(4):485-488
[26] Z. Wang, X. Hu. Energy-aware variable partitioning and instruction schedulingfor multibank memory architectures [J]. ACM Tractions on Design Automation ofElectronic Systems,2005,10(2):369-388
[27] M.Lajolo, A.Raghunathan, S. Dey, et al. Efficient power estimation techniquesfor hw/sw systems [C]. IEEE Alessandro Volta Memorial Workshop on Low-PowerDesign,1999:191-199
[28] A. Sama, M. Balakrishman, J. F. M. Theeuwen. Speeding up power estimationof embedded software [C]. Proceeding of Low Power Electronics and Design,2000:191-196
[29] E. Y. Chung, L. Benini, D. Micheli G. Sourec code transformation based onsoftware cost analysis [C]. Proceeding of ISSS,2001:153-158
[30]钱贾敏，王力生．基于复杂度的嵌入式软件功耗模型[J]．单片机与嵌入式系统应用，2004，9：18-20
[31] T. K. Tan, A. Raghunathan, N. K. Jha. Software Architectural Transformations:A New Approach to Low Energy Embedded Software [C]. Proceedings of theConference on Design, Automation and Test in Europe,2007:467-484
[32] Y. Wang, X. H. Shen, L. Zhang. Systemic model for software cost analyzing andoptimizing [C]. IEEE3rdInternational Conference on Communication Software andNetworks,2011:569-573
[33]胡俊，郭兵，沈艳等．基于灰色层次模型的嵌入式软件能耗优化方法评价研究[J]．四川大学学报(自然科学版)，2011，48(5)：1029-1035
[34]熊悦．面向对象软件能耗模型与能耗评估环境研究[D]：[博士学位论文]．安徽：中国科学技术大学，2003
[35] L. A. Zadeh. Fuzzy sets [J]. Information and Control,1965,8:338-353
[36] L. A. Zadeh. Outline of a new approach to the analysis complex systems anddecision processe [J]. IEEE Transaction Systems Man Cyberm.,1973, SMC3:28-44
[37] S. L. Chen, J. R. Wu. SR-convergence theory in fuzzy lattices [J]. InformationScience,2000,125:233-247
[38] S. L. Chen, S. T. Chen, X. G. Wang. σ-convergence theory in fuzzy lattices [J].Information Science,2004,165(2):45-58
[39]王国俊．模糊推理与模糊逻辑[J]．系统工程学报，1998，13(2):1-16
[40]朱光兴，杨德礼．模糊判断矩阵排序向量的确定方法研究[J]．模糊系统与数学，2004，18(2)：73-82
[41]陈水利，李敬功，王向公．模糊集理论及应用[M]．北京：科学出版社，2011
[42]汪培庄．模糊集与随机集落影[M]．北京：北京示范大学出版社，1985
[43]尤晋元，史美林．Windows操作系统原理[M]．北京：机械工业出版社，2001
[44]陈静华，陈迪平，徐勇军等．面向芯核设计的功耗层次化管理策略[J]．计算机辅助设计与图形学学报，2005，17(5)：1079-1084
[45]车健．嵌入式系统中低能耗设计[J]．电子测量技术，2005(3)：16-18
[46] L. Benini, A. Bogliolo, D. Micheli G. A survey of design techniques forsystem-level dynamic power management [J]. IEEE Transactions on VLSI Systems,2000,8(3):299-315
[47]吴琦．嵌入式操作系统功耗管理技术研究[D]：[博士学位论文]．成都：电子科技大学计算机科学与工程学院，2006
[48] F. Douglis, P. Krishnan, B. marsh. Thwarting the power-hungry disk [C].Proceeding of the Usenix Tenical Conference,1994:292-306
[49] P. Greenawalt. Modeling power management for hard disk [C]. Proceeding onModeling, Analysis, and Simulation of Computer and Telecommunication Systems,1994:62-66
[50] K. Li, R. Kumpf, P. Horton, et al. A quantitative analysis of disk drive powermanagement in portable computers [C]. Proceeding of the Usenix TechnicalConference,1994:279-291
[51] D. Helmbold, D. Long, B. Sherrod. A dynamic disk spin-down technique formobile computing [C]. Proceeding of the2ndAnnual ACM International Conferenceon Mobile Computing and NetWorking,1996:130-142
[52] F. Douglis, P. Krishnan, B. Bershad. Adaptive disk spin-down policies formobile computers [C]. Proceeding of2ndUsenix Symp. on Mobile andLocation-Independent Computing(MOBLIC),1995
[53] Y. Lu, G. Micheli. Adaptive hard disk power management on personalcomputers [C]. Proceeding of IEEE Great Lakes Symposium on VLSI,1999:50-53
[54] P. krishnan, P. Long, J. Vitter. Adaptive disk spindown via optimal rent-to-buy inprobabilistic environments [C]. Proceeding of International Conference of MachineLearning,1995:322-330
[55]赵霞，陈向群，郭耀等．操作系统电源管理研究进展[J]．计算机研究与发展，2008，45(5)：817-822
[56]戚隆宁，张哲，黄少珉．多任务下I/O设备的动态功耗管理[J]．中国工程科学，2008，10(2)：60-65
[57]戚隆宁，张哲，胡晨，卜爱国．基于MSR模型的动态功耗管理策略[J]．电路与系统学报，2006，11(5)：88-92
[58] M. Srivastava, Achnadrakasan R. Brodersne. Predictive Systme Shutdown andother architectura1techniuqes of energy efficient porgrammble computation [J]. IEEETransaction On VLSI System,1996,4(l):42-55
[59] E. Chung, L. Benini, G. D. Micheli. Dynamic power management usingadaptive learning trees [C]. Proceeding of International Conference onComputer-Aided Design,1999
[60] C. H. Hwang, A. Wu. A predictive system shutdown method for energy savingof event-driven computation [C]. Proceeding of International Conference ofComputer-Aided Design,1997:28-32
[61] C. Gniady, Y. Hu, Y. H. Lu. Program counter based techniques for dynamicpower management [C]. Proceeding of10thInternational Symposium on HighPerformance Computer Architecture,2004:24-35
[62]黄少珉，张哲，胡晨，戚隆宁．基于数据缓冲区动态分组的功耗管理策略[J]．电路与系统学报，200712(1)：26-32
[63]黄少珉，戚隆宁，杨军，胡晨．面向任务信息的层次化存储DPM策略[J]．中国工程科学，2010，12(2)：:83-89
[64]黄少珉，戚隆宁，杨军，胡晨．能耗期望的动态功耗管理预测性策略[J]．应用科学学报，2009，27(3)：272-276
[65]戚隆宁，胡晨，张哲，卜爱国．基于时间期望表的DPM预测策略[J]．电路与系统学报，2007，12(2)：89-94
[66]黄少珉，戚隆宁，胡晨．基于软件测量与控制的动态功耗管理预测方案[J]．仪器仪表学报，2006，27(6)：954-956
[67]卜爱国，刘昊，胡晨，王超．基于AR模型的PCF动态电源管理预测策略[J]．应用科学学报，2005(5)：483-488
[68]何可佳．基于概率的自适应学习预测策略[J]．计算机工程，2010，36(10)：215-217
[69]李顺，曾超，李军．一种自适应的动态功耗管理预测策略[J]．信息与电子工程，2010(4)：485-488
[70]荆元利，樊晓桠，高德远，沈戈，胡剑．一种数字信号处理器的动态功耗管理方案[J]．微电子学与计算机，2003(9)：60-63
[71] L. Benini, A. Bogliolo, G. Paleologo, et al. Policy optimization for dynamicpower management [J]. IEEE Transaction on Computer-Aided Design of IntegratedCircuits and Systems,1999,18(6):813-833
[72] Z. Ren, B. H. Krogh, R. Marculesu. Hierarchical adaptive dynamic powermanagement [J]. IEEE Trans. Computers,2005,54(4):409-420
[73] Q. Qiu, M. Pedram. Dynamic power management based on continuous-timemarkov decision processes [C]. Proceeding of Design Automation Conference,1999:555-561
[74] Q. Qiu, M. Pedram. Dynamic power management of complex systems usinggeneralized stochastic petri nets [C]. Proc. of Design Automation Conference,2000:352-356
[75] T. Simunic, L. Benini, G. D. Micheli. Event-driven power management [J].IEEE Trans. On Computer-Aided Design of Integrated Circuits and Systems,2001,20(7):840-856
[76]江琦，奚宏生，殷保群．动态电源管理的随机切换模型与在线优化[J]．自动化学报，2007，33(1)：66-71
[77]郑倩．基于随机模型的动态电源管理[D]：[硕士学位论文]．哈尔滨：哈尔滨工业大学航天学院，2010
[78] M. Weiser, B. Welch, A. Demers, S.Shenker. Scheduling for reduced CPUenergy [C]. Proceeding of the1stUSENIX Symp. on Operating Systems Design andImplementation,1994:13-23
[79] J. R. Lorch. Operating systems techniques for reducing processor energyconsumption [D]:[Ph.D. Thesis]. Berkeley: University of California,2001
[80] D. Mosse, H. Aydin, et al. Compiler-Assisted dynamic power-aware schedulingfor real-time applications [C]. Proceeding of the Workshop on Compilers andOperating Systems for Low-Power (COLP2000),2000:194-203
[81] D. K. Shin, J. H. Kim, S. S. Lee. Intra-Task volgage scheduling for low-energyhard real-time applications [J]. IEEE Design&Test of Computers,2001,18(2):20-30
[82] H. Saputra, M. Kandemir, et al. Energy-Consicious compilation based onvoltage scaling [C]. Proceeding of the Sigplan Joint Conf. on Languages, Compilers,and Tools for Embedded Systems and Software and Compilers for Embedded Systems,2002:1-10
[83] F. Gruian. Hare real-time scheduling for low-energy using stochastic data andDVS processors [C]. Proceeding of the Int’1Symp. on Low-Power Electronics andDesign,2001:46-51
[84] A. Azevedo, I. Issenin, et al. Profile-Based dynamic voltage scheduling usingprogram checkpoints [C]. Proceeding of the Design, Automation and Test in EuropeConference (DATE),2002:168-178
[85] N. AbouGhazaleh, D. Mosse, et al. Collaborative operating system and compilerpower management for real-time applications [C]. Proceeding of the Real-timeTechnology and Application Symp,2003:133-143
[86] C. H. Hsu, U. Kremer. The design, implementation, and evaluation of acompiler algorithm for CPU energy reduction [C]. Proceeding of the ACM Sigplan2003Conference on Programming language Design and Implementation,2003:38-48
[87]卜爱国．针对离散电压处理器的动态电压调节策略[J]．计算机应用研究，2011，28(5)：1786-1788
[88]王彪，王小鸽．功耗管理中的动态电压调整综述[J]．计算机应用研究，2007，24(8)：8-12
[89]闫家年，陈文光，郑纬民．面向结构体数据布局优化的高效内存管理[J]．清华大学学报（自然科学版），2011(1)：68-72
[90]李平勇，游磊．嵌入式操作系统μC／OS-Ⅱ的一种内存管理算法[J]．微电子学与计算机，2011(1)：98-101
[91]高超，韩锐，倪宏．嵌入式Linux平台内存管理方案[J]．小型微型计算机系统，2011(4)：614-618
[92]叶新栋，唐志强，涂时亮．RTOS动态分区内存管理机制的优化设计[J]．单片机与嵌入式系统应用，2009(9)：9-11
[93] G. Gao. Proactive power-aware cache management for mobile computingsystems [J]. Transactions on Computers,2002,51(6):608-621
[94] S. Kim, N. Viaykrishnan, M. Kandemir, et al. Energy-efficient instruction cacheusing page-based placement [C]. Proceeding of International Conference onCompilers, Architectures, and Systhesis for Embedded Systems,2001
[95] M. Kandemir, N. Vijaykrishnan, M. Irwin, et al. Influence of compileroptimizations on system power [C]. Proceeding of the37thConference on DesignAutomation,2000:304-307
[96] S. Udayakumaran, A. Dominguez, R. Barua. Dynamic allocation for scratch-padmemory using compile-time decisions [J]. ACM Transactions on EmbeddedComputing Systems (TECS),2006,5(2):1-34
[97] L. Li, L. Gao, J. Xue. Memory coloring: a compiler approach for automaticscratchpad memory management [C]. Proceeding of14thInternational Conference onParallel Architectures and Compilation Techniques2005:329-338
[98] S. Stenike, L. Wehmeyer, B. S. Lee, et al. Assigning program and data objects:to scratchpad for energy reduction [C]. Proceeding of Design, Automation and Test inEurope (DATE),2002:409-417
[99] Z. Wan, X. hu. Energy-aware variable partitioning and instruction schedulingfor multibank memory architectures [J]. ACM Transactions on Design Automation ofElectronic Systems,2005,10(2):369-388
[100] W. Zhang, J. S. Hu, V. Degalahal, et al. Compiler-directed Instruction cacheleakage optimization [C]. In Proc. of35thAnnual International Symposium onMicroarchitecture,2002:208-218
[101] A. R. Lebeck, X. Fan, H. Zeng, et al. Power aware page allocation [C].Proceeding of Architectural Support for Programming languages and OperatingSystems,2000:105-116
[102] H. Huang, P. Pillai, K. Shin. Design and implementation of power-aware virtualmemory [C]. Proceeding of USENIX Technical Conference,2003:5-5
[103] R. P. Dick, G. Lakshminarayana, A. Raghunathan, et al. Power analysis ofembedded operating systems [C]. Proceeding of the37thConference on Designautomation,2000:312-315
[104] A. Acquaviva, L. Benini, B. Ricco. Energy characterization of embeddedreal-time operating systems [J]. ACM SIGARCH Computer Architecture News,29(5),2001
[105] T. Cignetti, K. komarov, C. Ellis. Energy estimation tools for the palm [C].Proceeding of ACM MSWIM2000: Modeling, Analysis and Simulation of Wirelessand Mobile Systems,2000
[106] David P.Helmbold, Darrell D.E. Long, Tracey L. Sconyers, Bruce Sherrod.Adaptive disk spin-down for mobile computers [C]. Mobile Networks andApplications,2000:285-297
[107] C.Ruemmler, J. Wilkes. Unix disk access patterns [C]. Proceeding InternationalConference Usenix Technical,1993:405-420
[2]林闯，田源，姚敏．绿色网络和绿色评价：节能机制、模型和评价[J]．计算机学报，2011，34(4)：594-612
[3]王怀民，史佩昌，丁博，尹刚，史殿习．软件服务的在线演化[J].计算机学报，2011，2(34)：318-327
[4]刘欣洋．复杂网络环境下基于模型驱动的软件进化方法研究[D]：[硕士学位论文]．呼和浩特：内蒙古大学计算机学院，2011
[5]杜思奇.面向普适计算的软件动态更新[D]：[硕士学位论文]．上海：上海交通大学计算机科学与工程系，2011
[6] V. Tiwari, S. Malik, et al. Power analysis of embedded software: a first steptowards software power minimization [J]. IEEE Transactions on Very Large ScaleIntegration Systems (VLSI),1994,2(4):437-445
[7] V. Tiwari, S. Malik, A. Wolfe. Instruction Level power analysis andoptimization of software [J]. Journal of VLSL Signal Processing,1996(13):223-238
[8]陈嘉，董渊，杨阳，戴桂兰，王生原．基于指令聚类与指令调度的嵌入式软件功耗优化研究[J]．小型微型计算机系统，2006，1(27)：175-179
[9] C. Lee, J. K. Lee, T. T. Hwang. Compiler optimization on instruction schedulingfor low power [C]. ISSS2000,2000:20-22
[10]赵荣彩，唐志敏，张兆庆，Guang R. Gao．低功耗多线程编译优化技术[J]．软件学报，2002，13(6)：1123-1129
[11]赵荣彩，唐志敏，张兆庆，Guang R. Gao．软件流水的低功耗编译技术研究[J]．软件学报，2003，14(8)：1357-1363
[12]易会战，杨学军．有效的低功耗编译优化方法：部件使用局部化[J]．软件学报，2004，15(10)：1451-1460
[13] D. Vishal, C. P. Ravikumar. Software power optimization in an embeddedsystem [C]. Proceedings of the14th International Conference on VLSI Design,2000:254-259
[14] Kyu-won choi, Abhijit Chatterjee. Efficient instruction-level optimizationmethodology for low power embedded systems [C]. ACM Proceeding of theInternational Symposium on Synthesis,2001:147-152
[15] Brandolese Carlo. Source level estimation of energy consumption and executiontime of embedded software [C]. Proceedings of the11th EUROMICRO Conferenceon Digital System Design Architectures, Methods and Tools,2008:115-123
[16]廖海艳，范明明．源码级和算法级的功耗测试与优化[J]．单片机与嵌入式系统应用，2010，1：11-14
[17]罗刚，郭兵，沈艳，廖海艳，任磊．源程序级和算法级嵌入式软件功耗特性的分析与优化方法研究[J]．计算机学报，2009，32(9)：1869-1875
[18] T. Mukherjee, Q. H. Tang, C. Ziesman, et al. Software architecture for dynamicthermal management in datacenters [C]. Proceedings of the2nd InternationalConference on Communication Systems Software and Middle-ware,2007:1-11
[19] V. Swaminathan, K. Chakrabarty. Network flow techniques for dynamic voltagescaling in hard real-time systems [J]. IEEE Transactions on Computer-aided Design ofIntergrated Circuits and Systems,2004,23(10):1385-1398
[20]易会战，陈娟，杨学军，刘喆．基于语法树的实时动态电压调节低功耗算法[J]．软件学报，2005，16(10)：1726-1734
[21]吴琦，熊光泽．非平稳自相似业务下自适应动态功耗管理[J]．软件学报，2005，16(8)：1499-1505
[22]吴琦，熊光泽．基于随机决策模型的动态功耗管理策略研究[J]．计算机学报，2007，30(4)：622-628
[23] Q. Wu, G. Z. Xiong. Why simple timeout strategies work perfectly in practice[J]. Lecture Notes in Computer Science,2006(4221):555-564
[24] E. Y. Chung, L. Benini, et al. Dynamic power management for nonstationaryservice requests [J]. IEEE Transactions on Computers,2002,51(11):1345-1360
[25] S. Li, C. Zeng, et al. An adaptive dynamic power management predictive policy[J]. Information and Electronic Engineering,2010,8(4):485-488
[26] Z. Wang, X. Hu. Energy-aware variable partitioning and instruction schedulingfor multibank memory architectures [J]. ACM Tractions on Design Automation ofElectronic Systems,2005,10(2):369-388
[27] M.Lajolo, A.Raghunathan, S. Dey, et al. Efficient power estimation techniquesfor hw/sw systems [C]. IEEE Alessandro Volta Memorial Workshop on Low-PowerDesign,1999:191-199
[28] A. Sama, M. Balakrishman, J. F. M. Theeuwen. Speeding up power estimationof embedded software [C]. Proceeding of Low Power Electronics and Design,2000:191-196
[29] E. Y. Chung, L. Benini, D. Micheli G. Sourec code transformation based onsoftware cost analysis [C]. Proceeding of ISSS,2001:153-158
[30]钱贾敏，王力生．基于复杂度的嵌入式软件功耗模型[J]．单片机与嵌入式系统应用，2004，9：18-20
[31] T. K. Tan, A. Raghunathan, N. K. Jha. Software Architectural Transformations:A New Approach to Low Energy Embedded Software [C]. Proceedings of theConference on Design, Automation and Test in Europe,2007:467-484
[32] Y. Wang, X. H. Shen, L. Zhang. Systemic model for software cost analyzing andoptimizing [C]. IEEE3rdInternational Conference on Communication Software andNetworks,2011:569-573
[33]胡俊，郭兵，沈艳等．基于灰色层次模型的嵌入式软件能耗优化方法评价研究[J]．四川大学学报(自然科学版)，2011，48(5)：1029-1035
[34]熊悦．面向对象软件能耗模型与能耗评估环境研究[D]：[博士学位论文]．安徽：中国科学技术大学，2003
[35] L. A. Zadeh. Fuzzy sets [J]. Information and Control,1965,8:338-353
[36] L. A. Zadeh. Outline of a new approach to the analysis complex systems anddecision processe [J]. IEEE Transaction Systems Man Cyberm.,1973, SMC3:28-44
[37] S. L. Chen, J. R. Wu. SR-convergence theory in fuzzy lattices [J]. InformationScience,2000,125:233-247
[38] S. L. Chen, S. T. Chen, X. G. Wang. σ-convergence theory in fuzzy lattices [J].Information Science,2004,165(2):45-58
[39]王国俊．模糊推理与模糊逻辑[J]．系统工程学报，1998，13(2):1-16
[40]朱光兴，杨德礼．模糊判断矩阵排序向量的确定方法研究[J]．模糊系统与数学，2004，18(2)：73-82
[41]陈水利，李敬功，王向公．模糊集理论及应用[M]．北京：科学出版社，2011
[42]汪培庄．模糊集与随机集落影[M]．北京：北京示范大学出版社，1985
[43]尤晋元，史美林．Windows操作系统原理[M]．北京：机械工业出版社，2001
[44]陈静华，陈迪平，徐勇军等．面向芯核设计的功耗层次化管理策略[J]．计算机辅助设计与图形学学报，2005，17(5)：1079-1084
[45]车健．嵌入式系统中低能耗设计[J]．电子测量技术，2005(3)：16-18
[46] L. Benini, A. Bogliolo, D. Micheli G. A survey of design techniques forsystem-level dynamic power management [J]. IEEE Transactions on VLSI Systems,2000,8(3):299-315
[47]吴琦．嵌入式操作系统功耗管理技术研究[D]：[博士学位论文]．成都：电子科技大学计算机科学与工程学院，2006
[48] F. Douglis, P. Krishnan, B. marsh. Thwarting the power-hungry disk [C].Proceeding of the Usenix Tenical Conference,1994:292-306
[49] P. Greenawalt. Modeling power management for hard disk [C]. Proceeding onModeling, Analysis, and Simulation of Computer and Telecommunication Systems,1994:62-66
[50] K. Li, R. Kumpf, P. Horton, et al. A quantitative analysis of disk drive powermanagement in portable computers [C]. Proceeding of the Usenix TechnicalConference,1994:279-291
[51] D. Helmbold, D. Long, B. Sherrod. A dynamic disk spin-down technique formobile computing [C]. Proceeding of the2ndAnnual ACM International Conferenceon Mobile Computing and NetWorking,1996:130-142
[52] F. Douglis, P. Krishnan, B. Bershad. Adaptive disk spin-down policies formobile computers [C]. Proceeding of2ndUsenix Symp. on Mobile andLocation-Independent Computing(MOBLIC),1995
[53] Y. Lu, G. Micheli. Adaptive hard disk power management on personalcomputers [C]. Proceeding of IEEE Great Lakes Symposium on VLSI,1999:50-53
[54] P. krishnan, P. Long, J. Vitter. Adaptive disk spindown via optimal rent-to-buy inprobabilistic environments [C]. Proceeding of International Conference of MachineLearning,1995:322-330
[55]赵霞，陈向群，郭耀等．操作系统电源管理研究进展[J]．计算机研究与发展，2008，45(5)：817-822
[56]戚隆宁，张哲，黄少珉．多任务下I/O设备的动态功耗管理[J]．中国工程科学，2008，10(2)：60-65
[57]戚隆宁，张哲，胡晨，卜爱国．基于MSR模型的动态功耗管理策略[J]．电路与系统学报，2006，11(5)：88-92
[58] M. Srivastava, Achnadrakasan R. Brodersne. Predictive Systme Shutdown andother architectura1techniuqes of energy efficient porgrammble computation [J]. IEEETransaction On VLSI System,1996,4(l):42-55
[59] E. Chung, L. Benini, G. D. Micheli. Dynamic power management usingadaptive learning trees [C]. Proceeding of International Conference onComputer-Aided Design,1999
[60] C. H. Hwang, A. Wu. A predictive system shutdown method for energy savingof event-driven computation [C]. Proceeding of International Conference ofComputer-Aided Design,1997:28-32
[61] C. Gniady, Y. Hu, Y. H. Lu. Program counter based techniques for dynamicpower management [C]. Proceeding of10thInternational Symposium on HighPerformance Computer Architecture,2004:24-35
[62]黄少珉，张哲，胡晨，戚隆宁．基于数据缓冲区动态分组的功耗管理策略[J]．电路与系统学报，200712(1)：26-32
[63]黄少珉，戚隆宁，杨军，胡晨．面向任务信息的层次化存储DPM策略[J]．中国工程科学，2010，12(2)：:83-89
[64]黄少珉，戚隆宁，杨军，胡晨．能耗期望的动态功耗管理预测性策略[J]．应用科学学报，2009，27(3)：272-276
[65]戚隆宁，胡晨，张哲，卜爱国．基于时间期望表的DPM预测策略[J]．电路与系统学报，2007，12(2)：89-94
[66]黄少珉，戚隆宁，胡晨．基于软件测量与控制的动态功耗管理预测方案[J]．仪器仪表学报，2006，27(6)：954-956
[67]卜爱国，刘昊，胡晨，王超．基于AR模型的PCF动态电源管理预测策略[J]．应用科学学报，2005(5)：483-488
[68]何可佳．基于概率的自适应学习预测策略[J]．计算机工程，2010，36(10)：215-217
[69]李顺，曾超，李军．一种自适应的动态功耗管理预测策略[J]．信息与电子工程，2010(4)：485-488
[70]荆元利，樊晓桠，高德远，沈戈，胡剑．一种数字信号处理器的动态功耗管理方案[J]．微电子学与计算机，2003(9)：60-63
[71] L. Benini, A. Bogliolo, G. Paleologo, et al. Policy optimization for dynamicpower management [J]. IEEE Transaction on Computer-Aided Design of IntegratedCircuits and Systems,1999,18(6):813-833
[72] Z. Ren, B. H. Krogh, R. Marculesu. Hierarchical adaptive dynamic powermanagement [J]. IEEE Trans. Computers,2005,54(4):409-420
[73] Q. Qiu, M. Pedram. Dynamic power management based on continuous-timemarkov decision processes [C]. Proceeding of Design Automation Conference,1999:555-561
[74] Q. Qiu, M. Pedram. Dynamic power management of complex systems usinggeneralized stochastic petri nets [C]. Proc. of Design Automation Conference,2000:352-356
[75] T. Simunic, L. Benini, G. D. Micheli. Event-driven power management [J].IEEE Trans. On Computer-Aided Design of Integrated Circuits and Systems,2001,20(7):840-856
[76]江琦，奚宏生，殷保群．动态电源管理的随机切换模型与在线优化[J]．自动化学报，2007，33(1)：66-71
[77]郑倩．基于随机模型的动态电源管理[D]：[硕士学位论文]．哈尔滨：哈尔滨工业大学航天学院，2010
[78] M. Weiser, B. Welch, A. Demers, S.Shenker. Scheduling for reduced CPUenergy [C]. Proceeding of the1stUSENIX Symp. on Operating Systems Design andImplementation,1994:13-23
[79] J. R. Lorch. Operating systems techniques for reducing processor energyconsumption [D]:[Ph.D. Thesis]. Berkeley: University of California,2001
[80] D. Mosse, H. Aydin, et al. Compiler-Assisted dynamic power-aware schedulingfor real-time applications [C]. Proceeding of the Workshop on Compilers andOperating Systems for Low-Power (COLP2000),2000:194-203
[81] D. K. Shin, J. H. Kim, S. S. Lee. Intra-Task volgage scheduling for low-energyhard real-time applications [J]. IEEE Design&Test of Computers,2001,18(2):20-30
[82] H. Saputra, M. Kandemir, et al. Energy-Consicious compilation based onvoltage scaling [C]. Proceeding of the Sigplan Joint Conf. on Languages, Compilers,and Tools for Embedded Systems and Software and Compilers for Embedded Systems,2002:1-10
[83] F. Gruian. Hare real-time scheduling for low-energy using stochastic data andDVS processors [C]. Proceeding of the Int’1Symp. on Low-Power Electronics andDesign,2001:46-51
[84] A. Azevedo, I. Issenin, et al. Profile-Based dynamic voltage scheduling usingprogram checkpoints [C]. Proceeding of the Design, Automation and Test in EuropeConference (DATE),2002:168-178
[85] N. AbouGhazaleh, D. Mosse, et al. Collaborative operating system and compilerpower management for real-time applications [C]. Proceeding of the Real-timeTechnology and Application Symp,2003:133-143
[86] C. H. Hsu, U. Kremer. The design, implementation, and evaluation of acompiler algorithm for CPU energy reduction [C]. Proceeding of the ACM Sigplan2003Conference on Programming language Design and Implementation,2003:38-48
[87]卜爱国．针对离散电压处理器的动态电压调节策略[J]．计算机应用研究，2011，28(5)：1786-1788
[88]王彪，王小鸽．功耗管理中的动态电压调整综述[J]．计算机应用研究，2007，24(8)：8-12
[89]闫家年，陈文光，郑纬民．面向结构体数据布局优化的高效内存管理[J]．清华大学学报（自然科学版），2011(1)：68-72
[90]李平勇，游磊．嵌入式操作系统μC／OS-Ⅱ的一种内存管理算法[J]．微电子学与计算机，2011(1)：98-101
[91]高超，韩锐，倪宏．嵌入式Linux平台内存管理方案[J]．小型微型计算机系统，2011(4)：614-618
[92]叶新栋，唐志强，涂时亮．RTOS动态分区内存管理机制的优化设计[J]．单片机与嵌入式系统应用，2009(9)：9-11
[93] G. Gao. Proactive power-aware cache management for mobile computingsystems [J]. Transactions on Computers,2002,51(6):608-621
[94] S. Kim, N. Viaykrishnan, M. Kandemir, et al. Energy-efficient instruction cacheusing page-based placement [C]. Proceeding of International Conference onCompilers, Architectures, and Systhesis for Embedded Systems,2001
[95] M. Kandemir, N. Vijaykrishnan, M. Irwin, et al. Influence of compileroptimizations on system power [C]. Proceeding of the37thConference on DesignAutomation,2000:304-307
[96] S. Udayakumaran, A. Dominguez, R. Barua. Dynamic allocation for scratch-padmemory using compile-time decisions [J]. ACM Transactions on EmbeddedComputing Systems (TECS),2006,5(2):1-34
[97] L. Li, L. Gao, J. Xue. Memory coloring: a compiler approach for automaticscratchpad memory management [C]. Proceeding of14thInternational Conference onParallel Architectures and Compilation Techniques2005:329-338
[98] S. Stenike, L. Wehmeyer, B. S. Lee, et al. Assigning program and data objects:to scratchpad for energy reduction [C]. Proceeding of Design, Automation and Test inEurope (DATE),2002:409-417
[99] Z. Wan, X. hu. Energy-aware variable partitioning and instruction schedulingfor multibank memory architectures [J]. ACM Transactions on Design Automation ofElectronic Systems,2005,10(2):369-388
[100] W. Zhang, J. S. Hu, V. Degalahal, et al. Compiler-directed Instruction cacheleakage optimization [C]. In Proc. of35thAnnual International Symposium onMicroarchitecture,2002:208-218
[101] A. R. Lebeck, X. Fan, H. Zeng, et al. Power aware page allocation [C].Proceeding of Architectural Support for Programming languages and OperatingSystems,2000:105-116
[102] H. Huang, P. Pillai, K. Shin. Design and implementation of power-aware virtualmemory [C]. Proceeding of USENIX Technical Conference,2003:5-5
[103] R. P. Dick, G. Lakshminarayana, A. Raghunathan, et al. Power analysis ofembedded operating systems [C]. Proceeding of the37thConference on Designautomation,2000:312-315
[104] A. Acquaviva, L. Benini, B. Ricco. Energy characterization of embeddedreal-time operating systems [J]. ACM SIGARCH Computer Architecture News,29(5),2001
[105] T. Cignetti, K. komarov, C. Ellis. Energy estimation tools for the palm [C].Proceeding of ACM MSWIM2000: Modeling, Analysis and Simulation of Wirelessand Mobile Systems,2000
[106] David P.Helmbold, Darrell D.E. Long, Tracey L. Sconyers, Bruce Sherrod.Adaptive disk spin-down for mobile computers [C]. Mobile Networks andApplications,2000:285-297
[107] C.Ruemmler, J. Wilkes. Unix disk access patterns [C]. Proceeding InternationalConference Usenix Technical,1993:405-420