超级计算机系统实时节能控制技术的理论与实践研究

J4 ›› 2012, Vol. 34 ›› Issue (8): 24-31.

超级计算机系统实时节能控制技术的理论与实践研究

金士尧1，张冬松1，吴飞2

（1.并行与分布处理国防科技重点实验室，湖南长沙 410073;
2.上海工程技术大学电子电气工程学院,上海 201620）

收稿日期:2012-04-28 修回日期:2012-06-11 出版日期:2012-08-25 发布日期:2012-08-25
基金资助:
国家自然科学基金资助项目（61170284）；国家教育部博士点基金（20104307110005）；上海市教委科研创新项目（12ZZ182）

Research on the Theory and Practice of RealTime Energy Control Technology in Supercomputer Systems

JIN Shiyao1，ZHANG Dongsong1，WU Fei2

（1.National Laboratory of Parallel and Distributed Processing,Changsha 410073;
2.School of Electronic and Electrical Engineering,
Shanghai University of Engineering Science,Shanghai 201620,China）

Received:2012-04-28 Revised:2012-06-11 Online:2012-08-25 Published:2012-08-25

摘要/Abstract

摘要：

本文提出了一种解决超级计算机中系统级节能问题的方案，采用共享计算机资源实时任务执行概率和节点机安全切入和切出机制，实现了超级计算机系统中负载检测统计和预测以及节能安全决策。初步实验表明，本文提出的节能安全决策方法可以实现较多节能，其节能效果受限于具体的系统模型、开销模型和负载预测结果。

关键词: 超级计算机系统, 实时概率调度理论, 系统切入和切出节能控制

Abstract:

This paper presents a solution to the systemlevel energysaving program in supercomputer systems, which utilizes the realtime tasks executing the probability of shared computer resources and the safety mechanism to switch on and off the node machines, to achieve load detection statistics and forecasting of supercomputer systems, as well as energysaving safety decisions. The preliminary experiments show that the proposed energy security decisionmaking method can achieve more energy savings, and the effect of energy saving is limited by the specific system model, overhead model and the results of load forecasting.

Key words: supercomputer system;theory of realtime possibility scheduling;energyefficient control of switching on/off system

金士尧1，张冬松1，吴飞2. 超级计算机系统实时节能控制技术的理论与实践研究[J]. J4, 2012, 34(8): 24-31.

JIN Shiyao1，ZHANG Dongsong1，WU Fei2. Research on the Theory and Practice of RealTime Energy Control Technology in Supercomputer Systems[J]. J4, 2012, 34(8): 24-31.

参考文献

［1］Meuer H, Simon H,Strohmaier E, et al. TOP500 Supercomputer Site［EB/OL］. ［20111116］. http://www.top500. org.
［2］张云泉, 孙家昶. 2011年中国高性能计算机发展现状分析与展望［J］. 高性能计算机发展与应用, 2012, 38(1):28.
［3］卢宇彤. 面向高效能计算的大规模资源管理技术研究与实现［博士论文］［D］.长沙：国防科学技术大学，2009.
［4］Srikantaiah S,Kansal A,Zhao F.Energy Aware Consolidatation for Cloud Computing［C］∥Proc of the 2008 Conference on Power Aware Computing and Systems, 2008:10.
［5］Barroso L A, Hlzle U. The Case for EnergyProportional Computing［J］. Computer, 2007, 40(12):3337.
［6］Krishna C M, Shin K G. RealTime Systems［M］. McGrawHill,1997.
［7］金士尧. 共享资源网络的实时任务接纳控制方法［P］.中国，ZL200810144000.3, 20101201.
［8］叶新铭, 王斌. 基于时间特征的网络流量预测模型［J］. 计算机科学, 2005, 32(7):3437.
［9］Paxson V,Floyd S. Wilde Area Traffic:the Failure of Poisson Modeling［J］. IEEE/ACM Transaction on Networking, 1995, 3(3):226244.
［10］薛薇.SPSS统计分析方法及应用［M］. 北京:电子工业出版社,2009.
［11］SANG Aimin, Li Sanqi. A Predictability Analysis of Network Traffic［C］∥Proc of the 19th Annual Joint Conference of IEEE Computer and Communication Societies, 2000:342351.
［12］李超, 张冬松, 金士尧. 面向流媒体应用的实时可执行概率获取方法研究［J］.计算机应用研究,2010,27(9):34623465.
［13］IBM BladeCenter QS22 Systems［EB/OL］.［20111116］.http://www.ibm.com/developerworks/.
［14］2006年超级计算环境运行情况总结［EB/OL］.［20060501］ http://wenku.baidu.com/view/c94d12f7f61fb7360b4 c6589.html.