面向高效能计算的大规模资源管理技术研究与实现
|
摘要
高性能计算发展到今天,已经从单一地追求高性能转向致力于实现系统的高效能,提高系统的实际性能、可编程性、可移植性和健壮性,降低系统的开发、运行以及维护成本。然而,由于百千万亿次以上高性能计算机系统具有规模庞大、结构复杂和组成异构多样等特点,为了实现高效能目标,系统必须解决实际应用的持续性能难以提升、管理效率低、可靠性差、能源消耗巨大等多个挑战性问题。这些问题对高效能计算机系统的大规模资源管理系统在性能、功能和可扩展性等多个方面提出了很高的要求,大规模资源管理技术成为高效能计算机系统实现的一个重大挑战性技术问题。
论文以我们自行研制的可扩展共享存储(S2MP:Scalable Shared Memory Processing)体系结构的高性能计算机系统上的大规模资源管理系统实现为基础,以面向高效能大规模并行计算机系统的高效资源管理技术为主要研究内容,在资源管理模型、资源管理系统的可扩展技术、综合优化的调度机制、用户作业自动恢复的容错管理方法以及系统能耗管理技术等方面展开研究,本文主要工作和贡献如下:
1、提出了大规模并行计算机系统的深度资源信息模型DRIM,克服了传统资源管理系统所关注的资源对象粒度过粗和资源描述能力不足的问题,针对高效能计算系统的特点建立了实体模型、功能模型和应用模型,更加全面、准确地描述了计算资源、通信资源、存储资源、多模式应用等各方面的特征,并将资源对象之间的关系模型化,使得管理策略更有效,管理功能可扩展性更好,为大规模并行系统高效的作业调度与资源分配提供了有力支撑。
2、设计了动态层次式级联资源管理结构,提出了基于自组织方式的级联服务动态创建方法,优化了资源管理系统的通信协议,设计了轻载的传输协议来减少大规模资源管理开销,采用硬件通讯机制实现高效的控制消息传递,通过全局操作与综合优化实现大规模作业的快速加载,解决了资源管理系统的规模可扩展问题。采用基于构件的系统实现结构支持资源管理的功能扩展。在由2048个多核处理器构成的S2MP体系结构的系统上进行了资源管理系统的实现和测试,测试结果表明系统具有良好的可扩展性。
3、提出了基于综合优先级的调度策略,综合考虑系统的作业属性、资源属性和服务属性中的多个因素,提升了调度机制的灵活性和有效性;设计了可变深度的回填调度策略MC-Backfill,实现了根据队列实际状态对Backfill的深度和频度的动态调整,较好地协调了系统的公平性目标和高吞吐率目标的实现。系统测试表明,MC-Backfill策略可以在用户对作业执行时间估计不准确的情况下较好地减少作业平均等待时间,提高系统吞吐率。
4、建立了一种高性能计算系统的故障分布模型,提出了基于Checkpoint/Restart的作业容错执行时间模型;设计了面向可靠性的检查点周期选择算法和最优结点集合选择方法,增强了系统中作业运行的可靠性;实现了基于Checkpoint机制的作业自动容错,避免了系统运行过程中的人工干预,降低了系统的平均故障恢复时间,提高了系统的可用性。
5、结合系统级和应用级的能耗管理技术,从资源管理系统的角度研究了全系统能耗管理,设计了能耗约束条件下的资源分配方法进行系统级的结点能耗管理;提出了基于负反馈的两级能耗管理模型进行应用级的能耗管理,基于访存带宽和I/O带宽的利用率,采用线性控制和模糊控制相结合的方法动态调整并行应用线程和进程数目,适时将空闲处理器核关闭以节约系统能耗。并给出了对能耗控制管理有效性的测试和分析。
The technology trend in supercomputing is changing from purely pursuing the peak performance to comprehensively pursuing the high productivity. High productivity computing system (HPCS) aims to improve the programmability, portability, robustness of the system, and reduce the development, running and maintenance costs. However, due to the various features such as very large scale, complex and heterogeneous architecture, next-generation teraflops and petaflops systems face some vital challenges when aiming at implementing the high productivity target. Specifically, these challenges include how to improve the sustained performance, reliability, scalability, flexibility, and how to significantly reduce the power consumption during the overall design. Particularly, these challenges have become several critical research issues in large scale resource management system (RMS) of HPCS.
Our research work is based on the implementation of the large-scale resource management system for our own high performance computer system which has the Scalable Shared Memory Processing (S2MP) architecture. Focusing on the development of high productivity resource management system for large-scale parallel systems, in this thesis, we systematically investigate some key techniques in efficient resource model, scalable RMS architecture, optimized scheduling policy, fault-tolerance job management, and power management and other related techniques. The main contributions of this thesis are as follows:
1. A deep resource information model (DRIM) for the large-scale parallel computing system, has been proposed. DRIM not only addresses the disadvantage of the coarse grain resource definitions in traditional resource management systems, but also provides more comprehensive and realistic resource objects. Specifically, DRIM establishes entity model, function model and application model, which can accurately characterize the computing resources, communication resource, storage resource and different types of applications. DRIM also abstracts the relationship between the resources to make the management policy more effective and the management capability more viable. In a word, DRIM could provide powerful support for the job scheduling and resource allocation in RMS.
2. A dynamic cascade resource management architecture has been proposed to create the cascade services dynamically based on self-organization mode. A light-weight optimized transportation protocol has been designed to reduce the management overhead and optimize the communication performance of control messages. A fast job-launching mechanism has been presented by using low-level hardware communication mechanism and collective operations. These could improve the scalability of RMS. The component-based system architecture has been used to support the function scalability of RMS. MCRM, Multiple Case Resource Management system, has been realized for the system with S2MP architecture. The experiments on a S2MP system with 2048 processors show that MCRM has a better scalability.
3. An integrated-priority scheduling policy has been proposed, which considers various factors of job attributes, resource attributes and service attributes in system, it can promote the flexibility and efficiency of the scheduling mechanism. MC-backfill scheduling policy has been designed, which could adjust the backfill depth and frequency according to the status of the job queue. MC-backfill can not only improve system throughput, but also consider system fairness. The experiment results show that with MC-backfill policy, even in the case of inaccurate estimation of job running time by users, the average waiting time of jobs can decrease, and the throughput of system is improved.
4. A model for the fault-tolerance job running time using checkpoint/restart technique based on Weibull failure distribution model for high performance computing system, has been proposed. Algorithms for calculating the best checkpoint interval and selecting the best collection of processors have been designed to increase the reliability of the system. An automatic job recovery mechanism has been implemented for the S2MP system. With checkpoint, the jobs can recovery automatically when system failure occurs. This method can avoid manual intervention, reduce the average time of fault recovery and increase the availability of system.
5. Two approaches for power management has been proposed for the large-scale RMS. An algorithm for properly scheduling jobs and allocating resources under the constraints of system energy consumption has been presented as the system-level approach. A model of Feedback based Two-Level Power Management (FTLPM) has been presented as the application level approach, which can reduce the redundant parallelism in the applications to decrease the energy consumption. FTLPM combines the linear control model and fuzzy control model to control the concurrency of threads and processes according to the memory bandwidth of multi-core processor and I/O bandwidth of file system. The experiment results show the effectiveness of our approaches.
论文以我们自行研制的可扩展共享存储(S2MP:Scalable Shared Memory Processing)体系结构的高性能计算机系统上的大规模资源管理系统实现为基础,以面向高效能大规模并行计算机系统的高效资源管理技术为主要研究内容,在资源管理模型、资源管理系统的可扩展技术、综合优化的调度机制、用户作业自动恢复的容错管理方法以及系统能耗管理技术等方面展开研究,本文主要工作和贡献如下:
1、提出了大规模并行计算机系统的深度资源信息模型DRIM,克服了传统资源管理系统所关注的资源对象粒度过粗和资源描述能力不足的问题,针对高效能计算系统的特点建立了实体模型、功能模型和应用模型,更加全面、准确地描述了计算资源、通信资源、存储资源、多模式应用等各方面的特征,并将资源对象之间的关系模型化,使得管理策略更有效,管理功能可扩展性更好,为大规模并行系统高效的作业调度与资源分配提供了有力支撑。
2、设计了动态层次式级联资源管理结构,提出了基于自组织方式的级联服务动态创建方法,优化了资源管理系统的通信协议,设计了轻载的传输协议来减少大规模资源管理开销,采用硬件通讯机制实现高效的控制消息传递,通过全局操作与综合优化实现大规模作业的快速加载,解决了资源管理系统的规模可扩展问题。采用基于构件的系统实现结构支持资源管理的功能扩展。在由2048个多核处理器构成的S2MP体系结构的系统上进行了资源管理系统的实现和测试,测试结果表明系统具有良好的可扩展性。
3、提出了基于综合优先级的调度策略,综合考虑系统的作业属性、资源属性和服务属性中的多个因素,提升了调度机制的灵活性和有效性;设计了可变深度的回填调度策略MC-Backfill,实现了根据队列实际状态对Backfill的深度和频度的动态调整,较好地协调了系统的公平性目标和高吞吐率目标的实现。系统测试表明,MC-Backfill策略可以在用户对作业执行时间估计不准确的情况下较好地减少作业平均等待时间,提高系统吞吐率。
4、建立了一种高性能计算系统的故障分布模型,提出了基于Checkpoint/Restart的作业容错执行时间模型;设计了面向可靠性的检查点周期选择算法和最优结点集合选择方法,增强了系统中作业运行的可靠性;实现了基于Checkpoint机制的作业自动容错,避免了系统运行过程中的人工干预,降低了系统的平均故障恢复时间,提高了系统的可用性。
5、结合系统级和应用级的能耗管理技术,从资源管理系统的角度研究了全系统能耗管理,设计了能耗约束条件下的资源分配方法进行系统级的结点能耗管理;提出了基于负反馈的两级能耗管理模型进行应用级的能耗管理,基于访存带宽和I/O带宽的利用率,采用线性控制和模糊控制相结合的方法动态调整并行应用线程和进程数目,适时将空闲处理器核关闭以节约系统能耗。并给出了对能耗控制管理有效性的测试和分析。
The technology trend in supercomputing is changing from purely pursuing the peak performance to comprehensively pursuing the high productivity. High productivity computing system (HPCS) aims to improve the programmability, portability, robustness of the system, and reduce the development, running and maintenance costs. However, due to the various features such as very large scale, complex and heterogeneous architecture, next-generation teraflops and petaflops systems face some vital challenges when aiming at implementing the high productivity target. Specifically, these challenges include how to improve the sustained performance, reliability, scalability, flexibility, and how to significantly reduce the power consumption during the overall design. Particularly, these challenges have become several critical research issues in large scale resource management system (RMS) of HPCS.
Our research work is based on the implementation of the large-scale resource management system for our own high performance computer system which has the Scalable Shared Memory Processing (S2MP) architecture. Focusing on the development of high productivity resource management system for large-scale parallel systems, in this thesis, we systematically investigate some key techniques in efficient resource model, scalable RMS architecture, optimized scheduling policy, fault-tolerance job management, and power management and other related techniques. The main contributions of this thesis are as follows:
1. A deep resource information model (DRIM) for the large-scale parallel computing system, has been proposed. DRIM not only addresses the disadvantage of the coarse grain resource definitions in traditional resource management systems, but also provides more comprehensive and realistic resource objects. Specifically, DRIM establishes entity model, function model and application model, which can accurately characterize the computing resources, communication resource, storage resource and different types of applications. DRIM also abstracts the relationship between the resources to make the management policy more effective and the management capability more viable. In a word, DRIM could provide powerful support for the job scheduling and resource allocation in RMS.
2. A dynamic cascade resource management architecture has been proposed to create the cascade services dynamically based on self-organization mode. A light-weight optimized transportation protocol has been designed to reduce the management overhead and optimize the communication performance of control messages. A fast job-launching mechanism has been presented by using low-level hardware communication mechanism and collective operations. These could improve the scalability of RMS. The component-based system architecture has been used to support the function scalability of RMS. MCRM, Multiple Case Resource Management system, has been realized for the system with S2MP architecture. The experiments on a S2MP system with 2048 processors show that MCRM has a better scalability.
3. An integrated-priority scheduling policy has been proposed, which considers various factors of job attributes, resource attributes and service attributes in system, it can promote the flexibility and efficiency of the scheduling mechanism. MC-backfill scheduling policy has been designed, which could adjust the backfill depth and frequency according to the status of the job queue. MC-backfill can not only improve system throughput, but also consider system fairness. The experiment results show that with MC-backfill policy, even in the case of inaccurate estimation of job running time by users, the average waiting time of jobs can decrease, and the throughput of system is improved.
4. A model for the fault-tolerance job running time using checkpoint/restart technique based on Weibull failure distribution model for high performance computing system, has been proposed. Algorithms for calculating the best checkpoint interval and selecting the best collection of processors have been designed to increase the reliability of the system. An automatic job recovery mechanism has been implemented for the S2MP system. With checkpoint, the jobs can recovery automatically when system failure occurs. This method can avoid manual intervention, reduce the average time of fault recovery and increase the availability of system.
5. Two approaches for power management has been proposed for the large-scale RMS. An algorithm for properly scheduling jobs and allocating resources under the constraints of system energy consumption has been presented as the system-level approach. A model of Feedback based Two-Level Power Management (FTLPM) has been presented as the application level approach, which can reduce the redundant parallelism in the applications to decrease the energy consumption. FTLPM combines the linear control model and fuzzy control model to control the concurrency of threads and processes according to the memory bandwidth of multi-core processor and I/O bandwidth of file system. The experiment results show the effectiveness of our approaches.
引文
[1] H. S. Stone. High-Performance Computer Architecture (3rd Edition). 1993.
[2]周毓麟,沈隆钧.高性能计算的应用与战略地位.中国科学院院刊, 1999.3
[3] IBM. Computer science reaches historic breakthrough: Supercomputer performs 1,000 trillion operations per second. http://www.ibm.com/ibm/ideasformibm /us/roadrunner, 2008.
[4] PETAFLOP. http://www.petaflop.info/, 2009.
[5] PetaFLOPS Enabling Technologies and Applications. http://www.hq.nasa.gov/hpcc/petaflops/, 2006.
[6] K. Kennedy. Long-Term Research in High End Computing. The PITAC Report and Its Implications for the Petaflops Initiative. http://www.cs.rice.edu/~ken/ Presentations/Petaflops.pdf, 1999.
[7] J. Skolnick. Putting the Pathway back into Protein Folding. Proceedings of National Academic Science, 2005,2265-2266.
[8]杨学军.高性能计算机技术发展的思考.中国计算机科学大会,西安, 2008.
[9] Cary super company. CRAY XT3 Measured Balance. http://www.cray.com, 2006.
[10] W.J. Camp, J.L. Tomkins. The Design Specification and Initial Implementation of the Red Storm Architecture. Sandia National Laboratory, October 2003.
[11] The Message Passing Interface (MPI) Standard. http://www.mcs.anl.gov/mpi/, 2008.
[12] OpenMP Architecture Review Board. OpenMP Application Programming. http://www.openmp.org, May 2005.
[13] T. G. Mattson, Hillsboro, G. Henry, Beaverton. An Overview of the Intel TFLOPS. Intel Technology Journal, http://www.ai.mit.edu/projects/aries/course/ notes/ascii_red.pdf, 1998.
[14] K. Koch. How does ASCI actually complete multi-month 1000-processor milestone simulations. In Conference on High Speed Computing, 2002.
[15] T. Mudge. Power: a First Class Design Constraint for Future Architectures. In Proceeding of the 7th International Conference on High Performance Computing (HiPC 2000). Bangalore, India, 2000,215-224.
[16] X. Feng, R. Ge, et al. Power and Energy Profiling of Scientific Applicationson Distributed Systems. 19th International Parallel and Distributed Processing Symposium (IPDPS 05), Denver, CO, 2005.
[17] N.R.Adiga, G.Almasi, et al. An Overview of the Blue Gene/L Supercomputer. In Proceedings of the 2002 ACM/IEEE Conference. In Proceedings of the 2002 ACM/IEEE Conference on Supercomputing (SC'02). Baltimore, Maryland, USA, 2002.
[18] U. Weiser. Microprocessors: Bypass the power wall. In Intel Academic Forum (Keynote presentation). Barcelona, Spain. April 2004.
[19] R. Rajamony, R. Bianchini. Energy management for server clusters. 16th Annual ACM International Conference on Supercomputing, 2002.
[20] Top500. http://www.top500.org. 2008.
[21] High Productivity Computing Systems. http://www.highproductivity.org/.
[22] J. W. Manke, J. Wu. Data-Intensive System Benchmark Suite Analysis and Specification. Atlantic Aerospace Electronics Corp, 1999.
[23] PACC. http://www.darpa.mil/ipto/research/pacc. 2000.
[24] PCA. http://www.darpa.mil/ipto/programs/pca. 2004.
[25] O. Hassaine. Issues in Selecting a Job Management System. CPRE Engineering-HPC, 2002.
[26] Portable Batch System. http://www.openpbs.org, 2009.
[27] Condor. http://www.cs.wisc.edu/condor/, 2009.
[28] LSF. http://www.platform.com, 2009.
[29] IBM. LoadLeveler. www-1.ibm.com/servers/eserver/ecatalog/us/software/ 4051.html, 2009.
[30] POE User Guide. UCRL-WEB-201527. /http/computing.llnl/LCdocs/poe, 2009.
[31] SLURM Reference Manual. UCRL-WEB-201386 , Lawrence Livermore National Laboratory, http://www.computing.llnl/LCdocs/slurm, 2006.
[32] Slurm: Simple Linux Utility for Resource Management. UCRL-MA-147996 REV 3. http://www.llnl.gov/linux/slurm,2009.
[33] LLNL. LCRM(DPCS)Reference Manual. http://computing.llnl/LCdocs/dpcs, 2007.
[34] CHAOS: Linux from Livermore. UCRL-WEB-200040,http://computing.llnl/ LCdocs/chaos, 2005.
[35] S. IQBAL, R. GUPTA, A. Y. FANG. Planning Considerations for Job Scheduling in HPC Clusters. Reprinted from Dell Power Solutions, 2005.
[36] D. G. Feitelson. A survey of scheduling in multiprogrammed. Technical Report RC 19790, IBM Research Division, 1994.
[37] D. Perkovic, P. Keleher. Randomization, speculation, and adaptation in batch schedulers. In Proceedings of Supercomputing 2000 (SC2000), 2000.
[38] D. Jackson, Q. Snell, M. Clement. Core algorithms of the Maui scheduler. In Proceedings of the Workshop on Job Scheduling Strategies for Parallel Processing. Springer-Verlag, 2001.
[39] A. Mualem, D. G. Feitelson. Utilization, predictability workloads, and user runtime estimates in scheduling. IEEE Transactions on Parallel and Distributed Systems, 2001, 12(6) :29–543.
[40] B. G. Lawson, E. Smirni. Multiple-queue backfilling scheduling with priorities and reservations for parallel system. Job Scheduling Strategies for Parallel Processing, Springer Verlag, 2002, 2537: 72–87.
[41] P. Keleher, D. Zotkin, D. Perkovic. Attacking the bottlenecks in backfilling schedulers. Cluster Computing: Software Tools and Applications, 2000, 245–254.
[42] D. Talby, D. Feitelson. Supporting priorities and improving utilization of the IBM SP scheduler using slacks-based backfilling. In Proceedings of the 13th International Parallel Processing Symposium, 1999.
[43] AW Mu'alem, DG Feitelson. Utilization, predictability, workloads, and user runtime estimates in scheduling the IBM SP2 with backfilling. In IEEE Transactions on Parallel and Distributed Computing, 2001, 12: 529–543.
[44] H. Z. J. Skovira, W. Chan. The easy -loadleveler API project. In JSSPP, 1996, 41-47.
[45] D. Thain, T. Tannenbaum, M. Livny. Distributed Computing in Practice: The Condor Experience, Journal of Concurrency and Computation, 2004, 17(2/4):323-356.
[46] K. Ahmed, N. Werstiuk et al. Architecting and Enterprise Grid with Platform Enterprise Grid Orchestrator. TECHNICAL WHITEPAPER, Platform Computing, 2006.
[47] Sun Microsystems. Sun ONE Grid Engine Administration and User's Guide. http://amsafsl.itep.run/sge_user.pdf, 2002.
[48] R. Brightwell. System Software R&D at Sandia: To Red Storm and Beyond. Scalable Computing Systems Department Center for Computation, Computers, Information and Math Sandia National Laboratories, 2007.
[49] Y. Ding. Recovery-Oriented Computing: Main Techniques of BuildingMultitier Dependability. SEMINAR ON SELF-HEALING SYSTEMS,2004.
[50] D. Patterson. Recovery Oriented Computing: A New Research Agenda for a New Century. HPCA, 2002.
[51] A. Brown. Spheres of Undo: A Framework for Extending Undo. http://roc.cd.berkeley.edu/undo/, 2004.
[52] E. N. Elnozahy, L. Alvisi, et al. A survey of rollback-recovery protocols in message-passing. ACM Computing Surveys, 2002, 34(3): 375-408.
[53] J. S. Plank, K. Li, and M. A. Puening. Diskless checkpointing. IEEE Trans. Parallel Distributed Systems, 1998, 972-986.
[54] A. Beguelin, E. Seligman, and P. Stephan. Application level fault tolerance in heterogeneous networks of workstations. Journal of Parallel and Distributed Computing, 1997,147-155.
[55] Z. Chen, G. E. Fagg, et al. Fault tolerant high performance computing by a coding approach. ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP 2005), 2005, 213-223.
[56] D. Marques. Automatic application-level checkpointing for high performance computing systems. PhD thesis, Cornell University, 2006.
[57] J. S. Plank, Y. Chen, et al. Memory exclusion: optimizing the performance of checkpointing systems. Software Practice and Experience, 1999, 125-142.
[58] G. Bronevetsky, D. Marques, et al. C3: A system for automating application- level checkpointing of mpi program. In Proceedings of the 16th International Workshop on Languages and Compilers for Parallel Computing (LCPC), 2003.
[59] G. Bronevetsky, M. Schulz, et al. Application-level checkpointing for shared memory programs. In Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), 2004.
[60] G. Bronevetsky, K. Pingali, and P. Stodghill. Experimental evaluation of application-level checkpointing for openmp programs. In International Conference on Supercomputing(ICS), 2006.
[61] A. Geist, C. Engelmann. Development of naturally fault tolerant algorithms for computing on 100,000 processors. http://www.csm.ornl.gov/, 2002.
[62] G. Bosilca, Z. Chen, J. Langou, J. Dongarra. Recovery patterns for iterative methods in a parallel unstable environment. Technical Report UT-CS-04-538, University of Tennessee, Knoxville, Tennessee, USA, 2004.
[63] P. D. Hough, M. E. Goldsby, E. Walsh. Algorithm dependent fault tolerance for distributed computing. Technical Report SAND2000-8219, Sandia NationalLaboratories, Livermore, CA, 2002.
[64] IBM System Blue Gene/P Solution. http://www-03.ibm.com/systems/ deepcomputing/bluegene/, 2007.
[65] M. Warren, E. Weigle, W. Feng. High-density computing: A 240-node beowulf in one cubic mete. Proceedings of IEEE/ACM Supercomputing'02. Baltimore, Maryland, 2002.
[66] L. Harrison. Transmeta Crusoe. http://www.transmeta.com/, 2005.
[67] F. Bellosa. The Benefits of Event-Driven Energy Accounting in Power-Sensitive systems. 9th ACM SIGOPS European Workshop. Kolding, Denmark, 2000.
[68] C. Isci, M. Martonosi. Identifying program power phase behavior using power vectors. In Proceedings of the IEEE International Workshop on Workload Characterization, 2003.
[69] W.L.Bircher, J.Law, et al. Effective Use of Performance Monitoring Counters for Run-Time Prediction of Power. SIGDA:ACM Special Interest Group on Design Automation. New York, USA, 2004.
[70] J. Flinn. Managing Battery Lifetime with Energy-Aware Adaptation. ACM Transactions on Computer Systems. 2004, 137-179.
[71] H. Yang. Power-aware compilation techniques for high performance processors. University of Delaware, 2004.
[72] C. Hsu, W. Feng. A Power Aware Run Time System for High Performance Computing. Super Computing. USA, 2005.
[73] G. Chen, O. Ozturk, et al. Energy-aware code replication for improving reliability in embedded chip multiprocessors. 19th IEEE System on Chip Conference (SOCC-2006). 2006.
[74] Intel. Enhanced Intel SpeedStep Technology for the Intel Pentium M Processor. Microprocessors White Papers, http://whitepapers.silicon.com/, 2004.
[75] AMD. Featuring AMD Power Now Technology. http://www.amd.com/us-en /Processors/ProductInformation/ , 2002.
[76] R. Ce, X. Feng, K. W. Cameron. Performance-constrained Distributed DVS Scheduling for Scientific Applications on Power-aware Clusters. Proceedings of the 2005 ACM/IEEE conference on Supercomputing. Washington,DC, USA, IEEE Computer Society, 2005.
[77] V. Sharma, A. Thomas, et al. Power-aware QoS management in web servers. 24th Annual IEEE Real-Time Systems Symposium. Cancun, Mexico, IEEE, 2003.
[78] S. Gurumurthi. The Need for Temperature-Aware Storage Systems. Proceedings of the Intersociety Conference on Thermal and Thermomechanical Phenomena electronic Systems, 2006.
[79] G. Schulz. MAID 2.0:Energy Savings without Performance Compromises. http://www.storageio.com, 2008.
[80] SuperMicro. Green500. http://www.green500.org, 2008.
[81] K. W. Cameron, G. Rong, X. Feng. High-performance, power-aware distributed computing for scientific applications. Computer, 2005,38(11):40-47.
[82] D. Bhardwaj. HPC Systems and Models. Department of Computer Science & Engineering, Indian Institute of Technology, Delhi India, http://www.cse.iitd.ac.in/~dheerajb, 2003.
[83] D. McCaughan. Parallel Computing Models. HPC Analyst SHARCNET, University of Guelph, 2007.
[84] D. Galli. A Look to Parallel Computers. http://www.bo.infn.it/herab/calcolo/ HBparalook.html, 1996.
[85] Los Alamos National Laboratory. Roadrunner System Overview. http://www.lanl.gov/orgs/hpc/roadrunner/pdfs, 2008.
[86] J.K. Bennett, J.B. Carter, et al. Distributed Shared Memory: Experience with Munin. Department of Computer Science Rice University, 1992.
[87] J. HENNESSY. Cache-Coherent Distributed Shared Memory: Perspectives on Its Development and Future Challenges. PROCEEDINGS OF THE IEEE, 1999, 87(3).
[88] SGI Inc. SGI Products: Servers and Supercomputers. SGI Altix Family, http://www.sgi.com/products/servers/altix/, 2008.
[89] S. Scott. Thinking Ahead:Future Architectures from Cray. Chief Technology Officer, Cray Supercomputer Company, 2007.
[90]英特尔.多核处理器:迈向四核、超越无限.白皮书, 2007.
[91] T. Agarwal, A. Sharma, V. Kal'e. Topology-aware task mapping for reducing communication contention on large parallel machines. Proceedings 20th IEEE International Parallel & Distributed Processing Symposium, 2006.
[92] S. Saini, D. Talcott. Parallel I/O Performance Characterization of Columbia and NEC SX-8 Superclusters. 21th International Parallel and Distributed Processing Symposium(IPDPS 2007), California, USA, 2007.
[93] H. Stüben. The Cray T3D as a production machine at Konrad-Zuse-Zentrum Berlin. Springer Berlin / Heidelberg, 1995.
[94] I. Foster. Designing and Building Parallel Programs. Superscript Editorial Production Services, 1994.
[95] Argonne Laboratory. PMI Design Thoughts. http://wiki.mcs.anl.gov/mpich2/, 2008.
[96] Common Information Model (CIM) Standards. http://www.dmtf.org/standards /cim/, 2009.
[97] J. Moreira, M. Brutman, J. Castanos. Designing a Highly-Scalable Operating System: The Blue Gene/L Story. SC2006, IEEE, 2006.
[98] K. Davis, A. Hoisie1, et al. A Performance and Scalability Analysis of the BlueGene/L Architecture. Performance and Architecture Laboratory (PAL), 2007.
[99] A. Meijster. Scalability and BlueGene/L. Centre for High Performance Computing and Visualization. University of Groningen ,The Netherlands, 2005.
[100] IBM System. IBM SYSTEM BLUE GENE/P SOLUTION. Extending the limits of breakthrough science, 2008.
[101] K. Haskell, B. Jennings, L. Martinez. ASC Alliance Center Site Visit Sandia National Laboratories. Sandia National Laboratory, http://asc.llnl.gov/alliances/alliances_archive/2005_sandia.pdf, 2005.
[102] S. M. Kelly, et al. N-Way (CNW):An Implementation of the Catamount Light Weight Kernel Supporting N-cores Version 2.0. Scalable System Software Department Sandia National Laboratories, http://cfwebprod.sandia.gov/cfdocs/CCIM/, 2008.
[103] S. Kelly. The Design of the Red Storm High Performance Computer. Sandia National Laboratories, http://www.sandia.gov/~smkelly, 2008.
[104] M. Jette, D. Auble. SLURM Status Report, Supercomputing 2008. S&T Principal Directorate - Computation Directorate, Lawrence Livermore National Laboratory, 2008.
[105] M. Jette. Resource Management using SLURM. The 7th International Conference on Linux Clusters, 2006.
[106] M. Jette. High Scalability Resource Management with SLURM. S&T Principal Directorate - Computation Directorate, 2008.
[107] Platform. Achieving Performance and Scalability. Platform Documentation, http://www.slac.stanford.edu/comp/unix/package/lsf/currdoc/, 2008.
[108] E. Frachtenberg, F. Petrini, et al. Scalable Resource Management in High Performance Computers. IEEE International Conference on Cluster Computing (CLUSTER 2002), Chicago, USA, 2002.
[109] E. Frachtenberg, F. Petrini, et al. Opportunities and Challenges of Modern Communication Architectures: Case Study with QsNet. 18th International Parallel and Distributed Processing Symposium (IPDPS'04) - Workshop 8, 2004.
[110] M. L. Massie, B. N. Chun, D. E. Culler. The Ganglia Distributed Monitoring System: Design,Implementation, and Experience. Parallel Computing, 2003, 30(7):817-840.
[111]谢旻,邢座程,卢宇彤,廖湘科. NICHAL通讯软件接口的设计与实现.计算机研究与发展(增刊), 39(10), 2002.
[112]谢旻,卢宇彤,周恩强. PICH2-CMEX:可扩展消息传递接口实现技术, MPICH2-CMEX:可扩展消息传递接口实现技术.计算机工程与应用, 2008.
[113] R. Brightwell, A. B. Maccabe. Scalability Limitations of VIA-Based Technologies in Supporting MPI. Proceedings of the Fourth MPI Developer's and User's Conference, 2000.
[114] R. Brightwell. System Software R&D at Sandia: To Red Storm and Beyond. Scalable Computing Systems Department Center for Computation, Computers, Information and Math Sandia National Laboratories, 2007.
[115] J. Weinberg. Job Scheduling on Parallel Systems. University of California, San Diego, 2005.
[116] D. G. Feitelson. A survey of scheduling in multiprogrammed. Technical Report RC 19790, IBM Research Division, 1994.
[117] D. Perkovic, P. Keleher, Randomization. Speculation and adaptation in batch schedulers. In Proceedings of Supercomputing 2000 (SC2000), 2000.
[118] D. Jackson, Q. Snell, M. Clement. Core algorithms of the Maui scheduler. In Proceedings of the Workshop on Job Scheduling Strategies for Parallel Processing. Springer-Verlag, 2001.
[119] A. Mualem, G. Feitelson. Utilization, predictability workloads, and user runtime estimates in scheduling. IEEE Transactions on Parallel and Distributed Systems, 2001,12(6):529–543.
[120] B. G. Lawson, E. Smirni. Multiple-queue backfilling scheduling with priorities and reservations for parallel system. Job Scheduling Strategies for Parallel Processing, Springer Verlag, 2002,2537:72–87.
[121] P. Keleher, D. Zotkin, D. Perkovic. Attacking the bottlenecks in backfilling schedulers. Cluster Computing: The Journal of Networks, Software Tools and Applications, 2000, 245–254.
[122] D. Talby, D. Feitelson. Supporting priorities and improving utilization of theIBM SP scheduler using slacks-based backfilling. In Proceedings of the 13th International Parallel Processing Symposium, 1999.
[123] S. Chiang, M. K. Vernon. Class-Partitioning Job Scheduling for Large-Scale Parallel Systems. Proceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA’04), Las Vegas, 2004.
[124] S. Srinivasan, R. Kettimuthu, et al. Characterization of Backfilling Strategies for Parallel Job Scheduling. Department of Computer and Information Science The Ohio State University, 2002.
[125] OSU, OSU Collective Benchmark. http://nowlab.cse.ohio-state.edu/, 2007.
[126] LLNL, IOR Benchmark. http://www.llnl.gov/asci/purple/benchmarks/ limited/ior/, 2008.
[127] J. Gustafson. Reevaluating Amdahl's Law. Communications of the ACM 1988.
[128] D. Feitelson. Parallel Workloads Archive. http://www.cs.huji.ac.il/labs/ parallel/workload/, 2008.
[129] BlueGene/L. http://www.llnl.gov/asc/computing_resources/bluegenel/, 2008.
[130] F. Petrini. Scaling to Thousands of Processors with Buffered Coscheduling. In Scaling to New Heights Workshop, 2002.
[131] W. Gropp, E. Lusk. Fault Tolerance in MPI Programs, Proceedings of the Cluster Computing and Grid Systems Conference. Proceedings of the Cluster Computing and Grid Systems Conference, 2002.
[132] E. Dubrova. Fault tolerant design: an introduction. Kluwer Academic Publishers, 2006.
[133] E. N. Elnozahy, L. Alvisi, et al. A survey of rollback-recovery protocols in message-passing systems. ACM Computing Surveys, 2002.
[134] J. S. Plank, M. G. Thomason. Processor allocation and checkpoint interval selection in cluster computing systems. Journal of Parallel and Distributed Computing, 2001, 1570-1590.
[135] A. Oliner, R. K. Sahoo, J. E. Moreira, M. Gupta, A. Sivasubramaniam. Fault-aware job scheduling for BlueGene/L systems. In Proceedings of the 18th International Parallel and Distributed Processing Symposium, 2004,64-73.
[136] R. K. Sahoo, A. J. Oliner, et al. Critical event prediction for proactive management in large-scale computer clusters. In Proceedings of the ACM SIGKDD, Intl. Conf. on Knowledge Discovery Data Mining, 2003,426-435.
[137] R. K. Sahoo, I. Rish, et al. Autonomic computing features for large-scaleserver management and control. In AIAC Workshop, IJCAI 2003, 2003.
[138] R. Vilalta, S. Ma. Predictive rare events in temporal domains. In Proceedings IEEE Conf. on Data Mining (ICDM.02), 2002, 474-481.
[139] A. B. Brown. A Recovery-Oriented Approach to Dependable Services: Repairing Past Errors with System-wide Undo. PhD thesis, Computer Science Division-University of California, Berkeley, 2003.
[140] N. Raju, G, Yudan Liu, C. B. Leangsuksunl. Reliability Analysis in HPC clusters. Proceedings of the High availability and Performance Computing ,HAPCW2006. Santa Fe, NM, USA, 2006.
[141]王松桂,程维虎,高旅端.概率论与数理统计.科学出版社, 2003.
[142]顾瑛.韦伯分布最佳线性无偏估计系数.电子工业出版社, 2004.
[143] E. Susan, L. Graham, Marc Snir, Cynthia A. Patterson. Geting up to speed:the future of supercomputing. The National Academies Press , Washington, D.C. http://www.nap.edu,2005.
[144] R. Bianchini, R. Rajamony. Energy Conservation in Clustered Servers. Computer, 2003, 36(12):44-49.
[145] J. Chase, R. Doyle. Energy Management for Server Clusters. Eighth Workshop on Hot Topics in Operating Systems, 2001.
[146] C. Lefurgy, K. Rajamani, et al. Energy Management for Commercial Servers. IBM Austin Research Lab, Computer, 2003,36(12): 39-43.
[147] Transmeta Corporation. http://www.transmeta.com/, 2009.
[148] Intel XScale Technology Overview. http://www.intel.com/design/intelxscale/. 2009.
[149] I. Kadayif, M. Kandemir, et al. Exploiting Processor Workload Heterogeneity for Reducing Energy Consumption in Chip Multiprocessors. Proceedings of Design, Automation and Test in Europe (DATE-04). Paris, France, 2004, 1158-1163.
[150] I. Kadayif, M. Kandemir, M. Karakoy. An Energy Saving Strategy Based on Adaptive Loop Parallelization. Proceedings of Design Automation Conference (DAC'02). New Orleans, Louisiana, USA, 2002, 195-200.
[151] I. Kadayif, M. Kandemir, U. Sezer. An Integer Linear Programming Based Approach for Parallelizing Applications in On-Chip Multiprocessors. Proceedings of the 39th IEEE/ACM Design Automation Conference (DAC-02). New Orleans, LA, USA, 2002, 703-708.
[152] R. C. Dorf. Modern Control Systems. 1996.
[153] M. A. Suleman, M. K. Qureshi, Y.N. Patt. Feedback-driven threading: power-efficient and high-performance execution of multi-threaded workloads onCMPs. ASPLOS’2008. 2008.
[154]诸静等.模糊控制原理与应用.机械工业出版社, 1995.
[155] NASA. NPB Benchmark. http://science.nas.nasa.gov/Software/NPB/, 2009.
[156]徐国荣等.多物质可压缩流体力学的数值欧拉方法.数值计算与计算机应用, 1998.
[157] B. J. Gentry, R. A. Martin, R. E. Daly. An Eulerian Differencing Method for Unsteady Compressible Flow Problems. Compute Physics, 1966.
[2]周毓麟,沈隆钧.高性能计算的应用与战略地位.中国科学院院刊, 1999.3
[3] IBM. Computer science reaches historic breakthrough: Supercomputer performs 1,000 trillion operations per second. http://www.ibm.com/ibm/ideasformibm /us/roadrunner, 2008.
[4] PETAFLOP. http://www.petaflop.info/, 2009.
[5] PetaFLOPS Enabling Technologies and Applications. http://www.hq.nasa.gov/hpcc/petaflops/, 2006.
[6] K. Kennedy. Long-Term Research in High End Computing. The PITAC Report and Its Implications for the Petaflops Initiative. http://www.cs.rice.edu/~ken/ Presentations/Petaflops.pdf, 1999.
[7] J. Skolnick. Putting the Pathway back into Protein Folding. Proceedings of National Academic Science, 2005,2265-2266.
[8]杨学军.高性能计算机技术发展的思考.中国计算机科学大会,西安, 2008.
[9] Cary super company. CRAY XT3 Measured Balance. http://www.cray.com, 2006.
[10] W.J. Camp, J.L. Tomkins. The Design Specification and Initial Implementation of the Red Storm Architecture. Sandia National Laboratory, October 2003.
[11] The Message Passing Interface (MPI) Standard. http://www.mcs.anl.gov/mpi/, 2008.
[12] OpenMP Architecture Review Board. OpenMP Application Programming. http://www.openmp.org, May 2005.
[13] T. G. Mattson, Hillsboro, G. Henry, Beaverton. An Overview of the Intel TFLOPS. Intel Technology Journal, http://www.ai.mit.edu/projects/aries/course/ notes/ascii_red.pdf, 1998.
[14] K. Koch. How does ASCI actually complete multi-month 1000-processor milestone simulations. In Conference on High Speed Computing, 2002.
[15] T. Mudge. Power: a First Class Design Constraint for Future Architectures. In Proceeding of the 7th International Conference on High Performance Computing (HiPC 2000). Bangalore, India, 2000,215-224.
[16] X. Feng, R. Ge, et al. Power and Energy Profiling of Scientific Applicationson Distributed Systems. 19th International Parallel and Distributed Processing Symposium (IPDPS 05), Denver, CO, 2005.
[17] N.R.Adiga, G.Almasi, et al. An Overview of the Blue Gene/L Supercomputer. In Proceedings of the 2002 ACM/IEEE Conference. In Proceedings of the 2002 ACM/IEEE Conference on Supercomputing (SC'02). Baltimore, Maryland, USA, 2002.
[18] U. Weiser. Microprocessors: Bypass the power wall. In Intel Academic Forum (Keynote presentation). Barcelona, Spain. April 2004.
[19] R. Rajamony, R. Bianchini. Energy management for server clusters. 16th Annual ACM International Conference on Supercomputing, 2002.
[20] Top500. http://www.top500.org. 2008.
[21] High Productivity Computing Systems. http://www.highproductivity.org/.
[22] J. W. Manke, J. Wu. Data-Intensive System Benchmark Suite Analysis and Specification. Atlantic Aerospace Electronics Corp, 1999.
[23] PACC. http://www.darpa.mil/ipto/research/pacc. 2000.
[24] PCA. http://www.darpa.mil/ipto/programs/pca. 2004.
[25] O. Hassaine. Issues in Selecting a Job Management System. CPRE Engineering-HPC, 2002.
[26] Portable Batch System. http://www.openpbs.org, 2009.
[27] Condor. http://www.cs.wisc.edu/condor/, 2009.
[28] LSF. http://www.platform.com, 2009.
[29] IBM. LoadLeveler. www-1.ibm.com/servers/eserver/ecatalog/us/software/ 4051.html, 2009.
[30] POE User Guide. UCRL-WEB-201527. /http/computing.llnl/LCdocs/poe, 2009.
[31] SLURM Reference Manual. UCRL-WEB-201386 , Lawrence Livermore National Laboratory, http://www.computing.llnl/LCdocs/slurm, 2006.
[32] Slurm: Simple Linux Utility for Resource Management. UCRL-MA-147996 REV 3. http://www.llnl.gov/linux/slurm,2009.
[33] LLNL. LCRM(DPCS)Reference Manual. http://computing.llnl/LCdocs/dpcs, 2007.
[34] CHAOS: Linux from Livermore. UCRL-WEB-200040,http://computing.llnl/ LCdocs/chaos, 2005.
[35] S. IQBAL, R. GUPTA, A. Y. FANG. Planning Considerations for Job Scheduling in HPC Clusters. Reprinted from Dell Power Solutions, 2005.
[36] D. G. Feitelson. A survey of scheduling in multiprogrammed. Technical Report RC 19790, IBM Research Division, 1994.
[37] D. Perkovic, P. Keleher. Randomization, speculation, and adaptation in batch schedulers. In Proceedings of Supercomputing 2000 (SC2000), 2000.
[38] D. Jackson, Q. Snell, M. Clement. Core algorithms of the Maui scheduler. In Proceedings of the Workshop on Job Scheduling Strategies for Parallel Processing. Springer-Verlag, 2001.
[39] A. Mualem, D. G. Feitelson. Utilization, predictability workloads, and user runtime estimates in scheduling. IEEE Transactions on Parallel and Distributed Systems, 2001, 12(6) :29–543.
[40] B. G. Lawson, E. Smirni. Multiple-queue backfilling scheduling with priorities and reservations for parallel system. Job Scheduling Strategies for Parallel Processing, Springer Verlag, 2002, 2537: 72–87.
[41] P. Keleher, D. Zotkin, D. Perkovic. Attacking the bottlenecks in backfilling schedulers. Cluster Computing: Software Tools and Applications, 2000, 245–254.
[42] D. Talby, D. Feitelson. Supporting priorities and improving utilization of the IBM SP scheduler using slacks-based backfilling. In Proceedings of the 13th International Parallel Processing Symposium, 1999.
[43] AW Mu'alem, DG Feitelson. Utilization, predictability, workloads, and user runtime estimates in scheduling the IBM SP2 with backfilling. In IEEE Transactions on Parallel and Distributed Computing, 2001, 12: 529–543.
[44] H. Z. J. Skovira, W. Chan. The easy -loadleveler API project. In JSSPP, 1996, 41-47.
[45] D. Thain, T. Tannenbaum, M. Livny. Distributed Computing in Practice: The Condor Experience, Journal of Concurrency and Computation, 2004, 17(2/4):323-356.
[46] K. Ahmed, N. Werstiuk et al. Architecting and Enterprise Grid with Platform Enterprise Grid Orchestrator. TECHNICAL WHITEPAPER, Platform Computing, 2006.
[47] Sun Microsystems. Sun ONE Grid Engine Administration and User's Guide. http://amsafsl.itep.run/sge_user.pdf, 2002.
[48] R. Brightwell. System Software R&D at Sandia: To Red Storm and Beyond. Scalable Computing Systems Department Center for Computation, Computers, Information and Math Sandia National Laboratories, 2007.
[49] Y. Ding. Recovery-Oriented Computing: Main Techniques of BuildingMultitier Dependability. SEMINAR ON SELF-HEALING SYSTEMS,2004.
[50] D. Patterson. Recovery Oriented Computing: A New Research Agenda for a New Century. HPCA, 2002.
[51] A. Brown. Spheres of Undo: A Framework for Extending Undo. http://roc.cd.berkeley.edu/undo/, 2004.
[52] E. N. Elnozahy, L. Alvisi, et al. A survey of rollback-recovery protocols in message-passing. ACM Computing Surveys, 2002, 34(3): 375-408.
[53] J. S. Plank, K. Li, and M. A. Puening. Diskless checkpointing. IEEE Trans. Parallel Distributed Systems, 1998, 972-986.
[54] A. Beguelin, E. Seligman, and P. Stephan. Application level fault tolerance in heterogeneous networks of workstations. Journal of Parallel and Distributed Computing, 1997,147-155.
[55] Z. Chen, G. E. Fagg, et al. Fault tolerant high performance computing by a coding approach. ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP 2005), 2005, 213-223.
[56] D. Marques. Automatic application-level checkpointing for high performance computing systems. PhD thesis, Cornell University, 2006.
[57] J. S. Plank, Y. Chen, et al. Memory exclusion: optimizing the performance of checkpointing systems. Software Practice and Experience, 1999, 125-142.
[58] G. Bronevetsky, D. Marques, et al. C3: A system for automating application- level checkpointing of mpi program. In Proceedings of the 16th International Workshop on Languages and Compilers for Parallel Computing (LCPC), 2003.
[59] G. Bronevetsky, M. Schulz, et al. Application-level checkpointing for shared memory programs. In Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), 2004.
[60] G. Bronevetsky, K. Pingali, and P. Stodghill. Experimental evaluation of application-level checkpointing for openmp programs. In International Conference on Supercomputing(ICS), 2006.
[61] A. Geist, C. Engelmann. Development of naturally fault tolerant algorithms for computing on 100,000 processors. http://www.csm.ornl.gov/, 2002.
[62] G. Bosilca, Z. Chen, J. Langou, J. Dongarra. Recovery patterns for iterative methods in a parallel unstable environment. Technical Report UT-CS-04-538, University of Tennessee, Knoxville, Tennessee, USA, 2004.
[63] P. D. Hough, M. E. Goldsby, E. Walsh. Algorithm dependent fault tolerance for distributed computing. Technical Report SAND2000-8219, Sandia NationalLaboratories, Livermore, CA, 2002.
[64] IBM System Blue Gene/P Solution. http://www-03.ibm.com/systems/ deepcomputing/bluegene/, 2007.
[65] M. Warren, E. Weigle, W. Feng. High-density computing: A 240-node beowulf in one cubic mete. Proceedings of IEEE/ACM Supercomputing'02. Baltimore, Maryland, 2002.
[66] L. Harrison. Transmeta Crusoe. http://www.transmeta.com/, 2005.
[67] F. Bellosa. The Benefits of Event-Driven Energy Accounting in Power-Sensitive systems. 9th ACM SIGOPS European Workshop. Kolding, Denmark, 2000.
[68] C. Isci, M. Martonosi. Identifying program power phase behavior using power vectors. In Proceedings of the IEEE International Workshop on Workload Characterization, 2003.
[69] W.L.Bircher, J.Law, et al. Effective Use of Performance Monitoring Counters for Run-Time Prediction of Power. SIGDA:ACM Special Interest Group on Design Automation. New York, USA, 2004.
[70] J. Flinn. Managing Battery Lifetime with Energy-Aware Adaptation. ACM Transactions on Computer Systems. 2004, 137-179.
[71] H. Yang. Power-aware compilation techniques for high performance processors. University of Delaware, 2004.
[72] C. Hsu, W. Feng. A Power Aware Run Time System for High Performance Computing. Super Computing. USA, 2005.
[73] G. Chen, O. Ozturk, et al. Energy-aware code replication for improving reliability in embedded chip multiprocessors. 19th IEEE System on Chip Conference (SOCC-2006). 2006.
[74] Intel. Enhanced Intel SpeedStep Technology for the Intel Pentium M Processor. Microprocessors White Papers, http://whitepapers.silicon.com/, 2004.
[75] AMD. Featuring AMD Power Now Technology. http://www.amd.com/us-en /Processors/ProductInformation/ , 2002.
[76] R. Ce, X. Feng, K. W. Cameron. Performance-constrained Distributed DVS Scheduling for Scientific Applications on Power-aware Clusters. Proceedings of the 2005 ACM/IEEE conference on Supercomputing. Washington,DC, USA, IEEE Computer Society, 2005.
[77] V. Sharma, A. Thomas, et al. Power-aware QoS management in web servers. 24th Annual IEEE Real-Time Systems Symposium. Cancun, Mexico, IEEE, 2003.
[78] S. Gurumurthi. The Need for Temperature-Aware Storage Systems. Proceedings of the Intersociety Conference on Thermal and Thermomechanical Phenomena electronic Systems, 2006.
[79] G. Schulz. MAID 2.0:Energy Savings without Performance Compromises. http://www.storageio.com, 2008.
[80] SuperMicro. Green500. http://www.green500.org, 2008.
[81] K. W. Cameron, G. Rong, X. Feng. High-performance, power-aware distributed computing for scientific applications. Computer, 2005,38(11):40-47.
[82] D. Bhardwaj. HPC Systems and Models. Department of Computer Science & Engineering, Indian Institute of Technology, Delhi India, http://www.cse.iitd.ac.in/~dheerajb, 2003.
[83] D. McCaughan. Parallel Computing Models. HPC Analyst SHARCNET, University of Guelph, 2007.
[84] D. Galli. A Look to Parallel Computers. http://www.bo.infn.it/herab/calcolo/ HBparalook.html, 1996.
[85] Los Alamos National Laboratory. Roadrunner System Overview. http://www.lanl.gov/orgs/hpc/roadrunner/pdfs, 2008.
[86] J.K. Bennett, J.B. Carter, et al. Distributed Shared Memory: Experience with Munin. Department of Computer Science Rice University, 1992.
[87] J. HENNESSY. Cache-Coherent Distributed Shared Memory: Perspectives on Its Development and Future Challenges. PROCEEDINGS OF THE IEEE, 1999, 87(3).
[88] SGI Inc. SGI Products: Servers and Supercomputers. SGI Altix Family, http://www.sgi.com/products/servers/altix/, 2008.
[89] S. Scott. Thinking Ahead:Future Architectures from Cray. Chief Technology Officer, Cray Supercomputer Company, 2007.
[90]英特尔.多核处理器:迈向四核、超越无限.白皮书, 2007.
[91] T. Agarwal, A. Sharma, V. Kal'e. Topology-aware task mapping for reducing communication contention on large parallel machines. Proceedings 20th IEEE International Parallel & Distributed Processing Symposium, 2006.
[92] S. Saini, D. Talcott. Parallel I/O Performance Characterization of Columbia and NEC SX-8 Superclusters. 21th International Parallel and Distributed Processing Symposium(IPDPS 2007), California, USA, 2007.
[93] H. Stüben. The Cray T3D as a production machine at Konrad-Zuse-Zentrum Berlin. Springer Berlin / Heidelberg, 1995.
[94] I. Foster. Designing and Building Parallel Programs. Superscript Editorial Production Services, 1994.
[95] Argonne Laboratory. PMI Design Thoughts. http://wiki.mcs.anl.gov/mpich2/, 2008.
[96] Common Information Model (CIM) Standards. http://www.dmtf.org/standards /cim/, 2009.
[97] J. Moreira, M. Brutman, J. Castanos. Designing a Highly-Scalable Operating System: The Blue Gene/L Story. SC2006, IEEE, 2006.
[98] K. Davis, A. Hoisie1, et al. A Performance and Scalability Analysis of the BlueGene/L Architecture. Performance and Architecture Laboratory (PAL), 2007.
[99] A. Meijster. Scalability and BlueGene/L. Centre for High Performance Computing and Visualization. University of Groningen ,The Netherlands, 2005.
[100] IBM System. IBM SYSTEM BLUE GENE/P SOLUTION. Extending the limits of breakthrough science, 2008.
[101] K. Haskell, B. Jennings, L. Martinez. ASC Alliance Center Site Visit Sandia National Laboratories. Sandia National Laboratory, http://asc.llnl.gov/alliances/alliances_archive/2005_sandia.pdf, 2005.
[102] S. M. Kelly, et al. N-Way (CNW):An Implementation of the Catamount Light Weight Kernel Supporting N-cores Version 2.0. Scalable System Software Department Sandia National Laboratories, http://cfwebprod.sandia.gov/cfdocs/CCIM/, 2008.
[103] S. Kelly. The Design of the Red Storm High Performance Computer. Sandia National Laboratories, http://www.sandia.gov/~smkelly, 2008.
[104] M. Jette, D. Auble. SLURM Status Report, Supercomputing 2008. S&T Principal Directorate - Computation Directorate, Lawrence Livermore National Laboratory, 2008.
[105] M. Jette. Resource Management using SLURM. The 7th International Conference on Linux Clusters, 2006.
[106] M. Jette. High Scalability Resource Management with SLURM. S&T Principal Directorate - Computation Directorate, 2008.
[107] Platform. Achieving Performance and Scalability. Platform Documentation, http://www.slac.stanford.edu/comp/unix/package/lsf/currdoc/, 2008.
[108] E. Frachtenberg, F. Petrini, et al. Scalable Resource Management in High Performance Computers. IEEE International Conference on Cluster Computing (CLUSTER 2002), Chicago, USA, 2002.
[109] E. Frachtenberg, F. Petrini, et al. Opportunities and Challenges of Modern Communication Architectures: Case Study with QsNet. 18th International Parallel and Distributed Processing Symposium (IPDPS'04) - Workshop 8, 2004.
[110] M. L. Massie, B. N. Chun, D. E. Culler. The Ganglia Distributed Monitoring System: Design,Implementation, and Experience. Parallel Computing, 2003, 30(7):817-840.
[111]谢旻,邢座程,卢宇彤,廖湘科. NICHAL通讯软件接口的设计与实现.计算机研究与发展(增刊), 39(10), 2002.
[112]谢旻,卢宇彤,周恩强. PICH2-CMEX:可扩展消息传递接口实现技术, MPICH2-CMEX:可扩展消息传递接口实现技术.计算机工程与应用, 2008.
[113] R. Brightwell, A. B. Maccabe. Scalability Limitations of VIA-Based Technologies in Supporting MPI. Proceedings of the Fourth MPI Developer's and User's Conference, 2000.
[114] R. Brightwell. System Software R&D at Sandia: To Red Storm and Beyond. Scalable Computing Systems Department Center for Computation, Computers, Information and Math Sandia National Laboratories, 2007.
[115] J. Weinberg. Job Scheduling on Parallel Systems. University of California, San Diego, 2005.
[116] D. G. Feitelson. A survey of scheduling in multiprogrammed. Technical Report RC 19790, IBM Research Division, 1994.
[117] D. Perkovic, P. Keleher, Randomization. Speculation and adaptation in batch schedulers. In Proceedings of Supercomputing 2000 (SC2000), 2000.
[118] D. Jackson, Q. Snell, M. Clement. Core algorithms of the Maui scheduler. In Proceedings of the Workshop on Job Scheduling Strategies for Parallel Processing. Springer-Verlag, 2001.
[119] A. Mualem, G. Feitelson. Utilization, predictability workloads, and user runtime estimates in scheduling. IEEE Transactions on Parallel and Distributed Systems, 2001,12(6):529–543.
[120] B. G. Lawson, E. Smirni. Multiple-queue backfilling scheduling with priorities and reservations for parallel system. Job Scheduling Strategies for Parallel Processing, Springer Verlag, 2002,2537:72–87.
[121] P. Keleher, D. Zotkin, D. Perkovic. Attacking the bottlenecks in backfilling schedulers. Cluster Computing: The Journal of Networks, Software Tools and Applications, 2000, 245–254.
[122] D. Talby, D. Feitelson. Supporting priorities and improving utilization of theIBM SP scheduler using slacks-based backfilling. In Proceedings of the 13th International Parallel Processing Symposium, 1999.
[123] S. Chiang, M. K. Vernon. Class-Partitioning Job Scheduling for Large-Scale Parallel Systems. Proceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA’04), Las Vegas, 2004.
[124] S. Srinivasan, R. Kettimuthu, et al. Characterization of Backfilling Strategies for Parallel Job Scheduling. Department of Computer and Information Science The Ohio State University, 2002.
[125] OSU, OSU Collective Benchmark. http://nowlab.cse.ohio-state.edu/, 2007.
[126] LLNL, IOR Benchmark. http://www.llnl.gov/asci/purple/benchmarks/ limited/ior/, 2008.
[127] J. Gustafson. Reevaluating Amdahl's Law. Communications of the ACM 1988.
[128] D. Feitelson. Parallel Workloads Archive. http://www.cs.huji.ac.il/labs/ parallel/workload/, 2008.
[129] BlueGene/L. http://www.llnl.gov/asc/computing_resources/bluegenel/, 2008.
[130] F. Petrini. Scaling to Thousands of Processors with Buffered Coscheduling. In Scaling to New Heights Workshop, 2002.
[131] W. Gropp, E. Lusk. Fault Tolerance in MPI Programs, Proceedings of the Cluster Computing and Grid Systems Conference. Proceedings of the Cluster Computing and Grid Systems Conference, 2002.
[132] E. Dubrova. Fault tolerant design: an introduction. Kluwer Academic Publishers, 2006.
[133] E. N. Elnozahy, L. Alvisi, et al. A survey of rollback-recovery protocols in message-passing systems. ACM Computing Surveys, 2002.
[134] J. S. Plank, M. G. Thomason. Processor allocation and checkpoint interval selection in cluster computing systems. Journal of Parallel and Distributed Computing, 2001, 1570-1590.
[135] A. Oliner, R. K. Sahoo, J. E. Moreira, M. Gupta, A. Sivasubramaniam. Fault-aware job scheduling for BlueGene/L systems. In Proceedings of the 18th International Parallel and Distributed Processing Symposium, 2004,64-73.
[136] R. K. Sahoo, A. J. Oliner, et al. Critical event prediction for proactive management in large-scale computer clusters. In Proceedings of the ACM SIGKDD, Intl. Conf. on Knowledge Discovery Data Mining, 2003,426-435.
[137] R. K. Sahoo, I. Rish, et al. Autonomic computing features for large-scaleserver management and control. In AIAC Workshop, IJCAI 2003, 2003.
[138] R. Vilalta, S. Ma. Predictive rare events in temporal domains. In Proceedings IEEE Conf. on Data Mining (ICDM.02), 2002, 474-481.
[139] A. B. Brown. A Recovery-Oriented Approach to Dependable Services: Repairing Past Errors with System-wide Undo. PhD thesis, Computer Science Division-University of California, Berkeley, 2003.
[140] N. Raju, G, Yudan Liu, C. B. Leangsuksunl. Reliability Analysis in HPC clusters. Proceedings of the High availability and Performance Computing ,HAPCW2006. Santa Fe, NM, USA, 2006.
[141]王松桂,程维虎,高旅端.概率论与数理统计.科学出版社, 2003.
[142]顾瑛.韦伯分布最佳线性无偏估计系数.电子工业出版社, 2004.
[143] E. Susan, L. Graham, Marc Snir, Cynthia A. Patterson. Geting up to speed:the future of supercomputing. The National Academies Press , Washington, D.C. http://www.nap.edu,2005.
[144] R. Bianchini, R. Rajamony. Energy Conservation in Clustered Servers. Computer, 2003, 36(12):44-49.
[145] J. Chase, R. Doyle. Energy Management for Server Clusters. Eighth Workshop on Hot Topics in Operating Systems, 2001.
[146] C. Lefurgy, K. Rajamani, et al. Energy Management for Commercial Servers. IBM Austin Research Lab, Computer, 2003,36(12): 39-43.
[147] Transmeta Corporation. http://www.transmeta.com/, 2009.
[148] Intel XScale Technology Overview. http://www.intel.com/design/intelxscale/. 2009.
[149] I. Kadayif, M. Kandemir, et al. Exploiting Processor Workload Heterogeneity for Reducing Energy Consumption in Chip Multiprocessors. Proceedings of Design, Automation and Test in Europe (DATE-04). Paris, France, 2004, 1158-1163.
[150] I. Kadayif, M. Kandemir, M. Karakoy. An Energy Saving Strategy Based on Adaptive Loop Parallelization. Proceedings of Design Automation Conference (DAC'02). New Orleans, Louisiana, USA, 2002, 195-200.
[151] I. Kadayif, M. Kandemir, U. Sezer. An Integer Linear Programming Based Approach for Parallelizing Applications in On-Chip Multiprocessors. Proceedings of the 39th IEEE/ACM Design Automation Conference (DAC-02). New Orleans, LA, USA, 2002, 703-708.
[152] R. C. Dorf. Modern Control Systems. 1996.
[153] M. A. Suleman, M. K. Qureshi, Y.N. Patt. Feedback-driven threading: power-efficient and high-performance execution of multi-threaded workloads onCMPs. ASPLOS’2008. 2008.
[154]诸静等.模糊控制原理与应用.机械工业出版社, 1995.
[155] NASA. NPB Benchmark. http://science.nas.nasa.gov/Software/NPB/, 2009.
[156]徐国荣等.多物质可压缩流体力学的数值欧拉方法.数值计算与计算机应用, 1998.
[157] B. J. Gentry, R. A. Martin, R. E. Daly. An Eulerian Differencing Method for Unsteady Compressible Flow Problems. Compute Physics, 1966.