期刊文献+

异构并行计算系统可扩展模型的实现 被引量:3

Implementation of Scalability Model for Heterogeneous Parallel Computing System
下载PDF
导出
摘要 效率和可扩展性是并行系统的2个重要的性能指标,虽然异构系统越来越普遍,但对其可扩展性的研究还很少。给出一种适合同构系统和异构系统的效率的定义,根据此定义对可扩展性进行分析,得出既适用于同构系统又适用于异构系统的等效率模型,并根据开销比得出在某一效率常数保持一致的情况下系统规模和工作负载应如何变化。实验结果表明该模型可以对效率和可扩展性进行较好的评测,并能预测算法的可扩展性。 Efficiency and scalability are two important concepts for analyzing the performance of parallel computing systems. Although heterogeneous systems have become more and more common, the research of the heterogeneous system's sealability returns very few back. This paper presents a definition of efficiency that can be fit for homogeneous and heterogeneous systems. According to this definition, it presents a scalability model that can be applied to both heterogeneous systems and homogeneous systems. Experimental results show that the model can estimate preferably the efficiency and scalability of parallel computing systems.
出处 《计算机工程》 CAS CSCD 北大核心 2009年第17期97-99,共3页 Computer Engineering
基金 山东省高等学校重点实验研究基金资助项目(2005-400) 曲阜师范大学校科研基金资助项目(XJ0734)
关键词 等效率 负载均衡 可扩展性 constant efficiency load balancing scalability
  • 相关文献

参考文献5

  • 1Sterling T, Becher D J, Savarese D. Beowulf." A Parallel Workstation for Scientific Computation[C]//Proc. of International Conference on Parallel Processing. Washington D. C., USA: [s. n.], 1995.
  • 2Chen Jun, Li Xiaornei. A Practical Scalability Metric[C]//Proc. of the HPC Asia'2000. Singapore: [s. n.], 2000.
  • 3Alan H K, Horace P F. Measuring Parallel Processor Performance[J]. Communications ofACM, 1990, 33(5): 539-543.
  • 4Barbosa J, Tavanes J, Padilha A J. Linear Algebra Algorithm in a Heterogeneous Cluster of Personal Computers[C]//Proc. of the 9th Heterogeneous Computing Workshop. Cancun, Mexico: [s. n.], 2000.
  • 5Clematis A, Corana A. Modeling Performance of Heterogeneous Parallel Computing Systems[J]. Parallel Computing, 1999, 25(9): 1131-1146.

同被引文献12

  • 1Kothe D B. Science Prospects and Benefits with Exascale Computing[R]. Oak Ridge National Laboratory, Tech. Rep.: ORNL/TM-2007/232, 2007.
  • 2Snir M, Bade D. A Framework for Measuring Supercomputer Productivity[J]. International Jourmd of High Performance Computing Applications, 2004, 18(4): 417-432.
  • 3Elnozahy E N, Alvisi L, Wang Yimin, et al. A Survey of Rollback- recovery Protocols in Message-passing Systems[J]. ACM Computing Surveys, 2002, 34(3): 375-408.
  • 4陈国良.并行计算:结构、算法、编程[M].北京:高等教育出版社,2004:88-94.
  • 5Grama A, Gupta A, Kumar V.Isoefficiency function: a scalability metric for parallel algorithms and architectures[J].IEEE Parallel & Distributed Technology, 1993,1 (3) : 12-21.
  • 6Sun X, Rover D.Scalability of parallel algorithm-machine combi-nations[J].IEEE Transactions on Parallel and Distributed System, 1994,5(6) : 599-613.
  • 7黄铠,徐志伟.可扩展并行计算:技术、结构与编程[M].北京:机械工业出版社,2000.
  • 8迟利华,刘杰,李晓梅,胡庆丰.并行算法与并行机相结合的可扩展性[J].计算机研究与发展,1999,36(1):47-51. 被引量:7
  • 9王与力,杨晓东.一种更有效的并行系统可扩展性模型[J].计算机学报,2001,24(1):84-90. 被引量:16
  • 10刘颖,吕方,王蕾,陈莉,崔慧敏,冯晓兵.异构并行编程模型研究与进展[J].软件学报,2014,25(7):1459-1475. 被引量:13

引证文献3

二级引证文献5

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