期刊文献+
任意字段
题名或关键词
题名
关键词
文摘
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息

基于LogP模型的MPI广播算法的性能分析与评价 被引量：1

Analysis and Evaluation of MPI Broadcast Algorithms Based on LogP Model
下载PDF
导出
摘要 LogP模型是描述并行系统的有效模型,广播是MPI并行编程中常用的通信类型.基于LogP模型,对常用的4种广播算法进行了理论分析,得出了各自的执行时间或上下界,从而有效地区分出了各个算法的利弊及最适合的并行计算环境.该分析具有通用性,同样可以用于MPI其他通信类型算法的分析、选择与设计. LogP model is a wide used model for parallel systems,and broadcast is a common collective operation in MPI programming. Based on LogP model,atheoretical analysis is made for four commonly used broadcast algorithms,and their exact execution time or low- er/upper bound is obtained. Then the advantages/disadvantages are shown, and the best suitable parallel environments are determined. The method can also be applied for analysis,selection and design of the other MPI collective communications.
作者 张凌晓 唐存东
机构地区 南阳理工学院计算机科学与技术系
出处 《华北水利水电学院学报》 2008年第1期92-95,共4页 North China Institute of Water Conservancy and Hydroelectric Power
关键词 LOGP模型 MPI广播算法 分析与评价 LogP model MPI broadcast algorithms analysis and evaluation
分类号 TP301.6 [自动化与计算机技术—计算机系统结构]
  • 相关文献

参考文献6

  • 1CULLER D E, KARP R M, PATTERSON D A, et al. LogP : Towards a Realistic Model of Parallel Computation [ C ] // Proc. Fourth ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, ACM SIGPLAN Notices 1993. 1 - 12.
  • 2BERNASCHI M, IANNELL G. Collective Communication Operations :Experimental Results vs. Theory [ J ]. Concurrency.. Practice and Experience, 1998,10 ( 5 ) :359 - 386.
  • 3Indiana University. LAM/MPI Parallel Computing [ EB/ O L ]. 2007 -02 - 14. http ://www. lam - mpi. org.
  • 4SATHISH S Vadhiyar, GRAHAM E Fagg, DONGARRA Jack. Automatically tuned collective communications[ C ] // Proceedings of SC99: High Performance Networking and Computing, 1999.
  • 5KARP R M,SCHAUSER K E. Optimal broadcast and summation in the LogP model [ C ] //Technical Report, UC Berkerly, 1992 : 1 - 39.
  • 6KIELMANN T,HOFMAN R F H,BAL H E,et al. MAG- PIE:MPI' s collective communication operations for clustered wide area systems [ C ]//ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, Atlanta,GA,1999:131 - 140.

同被引文献9

  • 1Thakur R,RabenseifnerR,Gropp W.Optimization of collective communication operations in MPICH[J].International Journal of High Performance Computing Applications,2005,(19)1:49-66.
  • 2Tu B,Zou M,Zhan J F,et al.Multi-core aware optimization for MPI collectives[C] // IEEE International Conference on Cluster Computing.Tsukuba,Japan:IEEE Press,2008:322-325.
  • 3Worringen J.Pipelining and overlapping for MPI collective operations[C] // Proceedings of the 28th Annual IEEE International Conference on Local Computer Networks.Bonn/Konigswinter,Gernany:IEEE Computer Society,2003::548-557.
  • 4Tipparaju V,Nieplocha J,Panda D.Fast collective operations using shared and remote memory access protocols on clusters[C] // Proceedings of the 17th International Symposium Parallel and Distributed Processing.Washington,USA:IEEE Computer Society,2003:84.
  • 5Thakur R,GroppW,Toonen B.Minimizing synchronization overhead in the implementation of MPI one-sided communication[C] //Proceeding of the 11th European PVM/MPI Users' Group Meeting.Springer,2004:57-67.
  • 6Bell C,Bonachea D,Nishtala R,et al.Optimizing bandwidth limited problems using one-sided communication and overlap[C] // Proceedings of the 20th International Conference on Parallel and Distributed Processing.Washington,USA:IEEE Computer Society,2006:84.
  • 7Kielmann T,Bal H E,Verstoep K.Fast measurement of LogP parameters for message passing platform[C] //Proceedings of the 15th IPDPS workshops on Parallel and Distributed Processing.Springer,2000,1 800:1 176-1 183.
  • 8高飞,刘青昆,向文,黄丹.MPICH标准通信模式下消息传递机制的研究[J].长春师范学院学报（自然科学版）,2007,26(5):65-69. 被引量：2
  • 9龙文光.CPU的流水线结构[J].四川师范大学学报（自然科学版）,2003,26(3):319-322. 被引量：1

引证文献1

二级引证文献1

华北水利水电学院学报
2008年 第1期

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