期刊文献+

一种基于超步执行的并行模拟平台研究与实现 被引量:3

Study and Implementation of Big-step Based Parallel Simulation Platform
下载PDF
导出
摘要 计算机系统的并行化趋势使得传统的串行模拟器正面临严峻的挑战,一方面由于执行速度和内存的限制无法模拟目标大规模并行系统;另一方面,无法充分利用已有并行系统的资源。因此,串行模拟器已经成为宿主机和目标系统之间的瓶颈。Hypersim是基于SMP集群平台的新型并行模拟器,运行于其上的主从式操作系统原型hyperOS为应用程序提供了运行环境,实验表明,该模拟平台可以正确地模拟一种新型的HPP高性能计算机体系结构。为了高效并行,hypersim中采用细粒度负载分配及高效的同步和通信机制,提出并采用超步执行技术来加速模拟器执行,结果显示hypersim在多进程情况下有一定的加速,在单进程多线程情况下有较好的加速比。 Simulation is facing the great challenge of computing capacity and memory demanding when used for studying extreme large and complex future systems, and sequential simulators have become a bottleneck since they cannot make full use of the abundant CPU and memory resources available on current parallel machines. Hypersim is a novel parallel simulator based on SMP cluster, and a master/client mode prototype operating hyperOS runs on it, results obtained from the experiments show that the simulation platform can simulate a novel HPP supercomputer architecture correctly. The techniques employed to parallelize the simulation was proposed in hypersim, including fine-grind load balance and high efficiency synchronization and communication..A new approach, big-step executing mechanism, was proposed and implemented to accelerate the parallel simulation. Results show that hypersim gets a better performance when running in multi-process mode, and can gain a sound speedup ratio in multi-thread mode.
出处 《系统仿真学报》 CAS CSCD 北大核心 2009年第15期4647-4653,4660,共8页 Journal of System Simulation
基金 国家高技术研究发展计划项目(2007AA01Z115)
关键词 并行模拟 超步执行 锁避免通信 HPP体系结构 parallel simulate big-step execution lock free communication HPP architecture
  • 相关文献

参考文献1

二级参考文献15

  • 1许建卫,陈明宇,包云岗.高带宽远程内存结构中的预取研究[J].计算机科学,2005,32(8):15-20. 被引量:2
  • 2Mendel Rosenblum,Stephen A.Herrod,Emmett Witchel et al. Complete Computer System Simulation:The SimOS Approach[J].IEEE Parallel and Distributed Technology, 1995 ;3(4) :34~43.
  • 3T Austin,E Larson. SimpleScalar:An Infrastructure for Computer System Modeling[J].IEEE Computer,2002;35(2):59~67.
  • 4Patrick Bohrer,Mootaz Elnozahy,Ahmed Gheith et al. Mambo-A Full System Simulator for the PowerPC Architecture[J].ACM SIGMETRICS Performance Evaluation Review, 2004; 31 (4): 8~12.
  • 5VMIPS Programmer's Manual.http:∥www.dgate.org/vmips/.
  • 6George Fankhauser,Christian Conrad,Eckart Zitzler et al.Topsy-A Teachable Operating System. http:∥www.tik.ee.ethz.ch/~topsy/.
  • 7SPEC CPU2000.http:∥www.specbench.org/osg/cpu2000/.
  • 8Synchronous DRAM.http:∥www.micronsemi.com/.
  • 9Cygwin User's Guide.http:∥cygwin.com/cygwin-ug-net/.
  • 10The Technology behind LynxOS v4.0's Linux ABI Compatibility.http:∥www.lynuxworks.com/products/.

共引文献3

同被引文献28

  • 1Ceze L,Strauss K,Almasi G et al. Full circle:SimulatingLinux clusters on Linux clusters//Proceedings of the 4th LCIInternational Conference on Linux Clusters:The HPC Revo-lution 2003. San Jose, USA, 2003:1-18.
  • 2Austin T,Larson E,Ernst D. Simplescalar:An infrastruc-ture for computer system modeling. IEEE Computer, 2002,35(2):59-67.
  • 3Rosenblum M, Herrod S, Witchel E, Gupta A. Completecomputer simulation:The SimOS approach. Parallel and Dis-tributed Technology, 1995,3(4):35-43.
  • 4Magnusson P S,Christensson M, Eskilson J et al. Simics:Afull system simulation platform. IEEE Computer, 2002,35(2):50-58.
  • 5Martin M,Sorin D J,Beckmann B M et al. Multifacetf sgeneral execution-driven multiprocessor simulator (GEMS)toolset. ACM SIGARCH Computer Architecture News,2005,33(4):92-99.
  • 6Binkert N L,Dreslinski R G,Hsu L R et al. The M5 simulator:Modeling networked systems. IEEE Micro, 2006,26(4):52-60.
  • 7Zheng G, KakulapatiG,Kale L. Bigsim:A parallel simulatorfor performance prediction of extremely large parallelmachines//Proceedings of the 18th International Parallel andDistributed Processing Symposium. Santa Fe, USA, 2004 :786-795.
  • 8Prakash S,Bagrodia R L. MPI-SIM:Using parallel simula-tion to evaluate mpi programs//Proceedings of the 30thConference on Winter Simulation(WSC 1998). Los Alamitos,USA, 1998:467-474.
  • 9Kale L V,Krishnan S. CHARM+ +:Parallel Programmingwith Message-Driven Objects. Parallel Programming usingC+ + . USA:MIT Press, 1996.
  • 10Joshua J Yi, Fenow-David J Lilja. Simulation of computer architectures: Simulators, benchmarks, methodologies and recommendations [J]. IEEE Transactions on Computers (S0018-9340), 2006, 55(3): 268-280.

引证文献3

二级引证文献8

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部