[1]田青,祝永志. SMP集群系统的可扩放性分析[J].计算机技术与发展,2017,27(06):95-98.
　TIAN Qing,ZHU Yong-zhi. Analysis on Scalability of SMP Cluster System[J].,2017,27(06):95-98.
点击复制

 SMP集群系统的可扩放性分析()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
27
期数:
2017年06期
页码:
95-98
栏目:
智能、算法、系统工程
出版日期:
2017-06-10

文章信息/Info

Title:
 Analysis on Scalability of SMP Cluster System
文章编号:
1673-629X（2017)06-0095-04
作者:
 田青; 祝永志
 曲阜师范大学 信息科学与工程学院
Author(s):
 TIAN Qing ; ZHU Yong-zhi
关键词:
 ; 并行计算; SMP集群; 可扩放性; 等效率
Keywords:
 ; parallel computing; SMP clusters; scalability; ISO-efficiency
分类号:
TP301
文献标志码:
A
摘要:
 随着并行计算技术的快速发展和SMP集群的普及,可扩放性已经成为并行应用程序设计和实现方面最重要的性能之一.但传统的可扩放性评价准则不能对SMP集群的可扩放性进行较精准的评价.为此,在分析SMP集群中处理器集合的特性和传统等效率模型并掌握其优缺点以及分析并行计算速度的基础上,给出了一种适合SMP集群系统效率的定义,并基于该定义提出了一种新的可扩放性评价准则(改进的等效率可扩放性评价准则).该新准则可用来评价并行算法和SMP集群相结合的可扩放性.为验证所提出评价准则的有效性,在集群平台上运行矩阵乘法程序进行了相关的扩放性实验.实验结果表明,所提出的新评价准则对算法和并行机的最优匹配有指导作用,同时有助于对并行算法的设计和改进.
Abstract:
 With the rapid development of parallel computing technology and the popularity of SMP clusters,scalability has become one of the most important performance for parallel application program design and implementation.However,the traditional scalability metric cannot evaluate the scalability of SMP cluster accurately.Therefore,the definition has been proposed to describe the efficiency of SMP cluster system based on the analysis of the property of the set of processors and the ISO-efficiency function in SMP cluster,understanding their merits and deficiencies,as well as the analysis of the parallel execution speed.Then a new metric,improved ISO-efficiency metric,has been proposed,which can be used to measure and evaluate the scalability of parallel algorithms and SMP cluster.The extension experiments have been carried out on the cluster platforms by running the program for the matrix multiplication algorithm in order to validate the effectiveness of the new evaluation criteria.Experimental results show that the new evaluation criterion has guided the optimal matching of the parallel application and the parallel machines.Thus it is helpful to the design and the improvement of parallel algorithms.

相似文献/References:

[1]龚向坚 邹腊梅 隆重.基于分布对象的虚拟网络实验系统设计与实现[J].计算机技术与发展,2010,(01):111.
　GONG Xiang-jian,ZOU La-mei,LONG Zhong.Design and Realization of Virtual Network Laboratory System Based on Distributing Object[J].,2010,(06):111.
[2]职为梅 王芳 范明 杨勇.并行环境下的同步异步PSO算法[J].计算机技术与发展,2009,(03):123.
　ZHI Wei-mei,WANG Fang,FAN Ming,et al.Synchronous and Asynchronous PSO Algorithm of Parallel Circumstance[J].,2009,(06):123.
[3]陈小飞 徐宏炳.基于网格的并行FFT计算研究[J].计算机技术与发展,2008,(03):67.
　CHEN Xiao-fei,XU Hong-bing.Research of Parallel FFT Computing Based on Grid[J].,2008,(06):67.
[4]陈荣征 李代平 黄健 秦昭晖.EBE-PCG算法在有限元并行计算中的应用研究[J].计算机技术与发展,2008,(03):232.
　CHEN Rong-zheng,LI Dai-ping,HUANG Jian,et al.Research on Application of EBE- PCG Algorithm in Parallel Computing of FEM[J].,2008,(06):232.
[5]王勇超 张璟 王新卫 马静.基于MPICH2的高性能计算集群系统研究[J].计算机技术与发展,2008,(09):101.
　WANG Yong-chao,ZHANG Jing,WANG Xin-wei,et al.Research of High Performance Cluster System Based on MPICH2[J].,2008,(06):101.
[6]王延玲 祝永志 郭静.基于HPJava集群系统的环境搭建与性能分析[J].计算机技术与发展,2008,(11):94.
　WANG Yan-ling,ZHU Yong-zhi,GUO Jing.Constructing and Perfomance Analysis of Cluster System Based on HPJava[J].,2008,(06):94.
[7]冯冲 颜廷华.基于线积分卷积算法的并行实现方法[J].计算机技术与发展,2008,(12):22.
　FENG Chong,YAN Ting-hua.A Parallel Algorithm for Line Integral Convolution[J].,2008,(06):22.
[8]张晓奇 张翌维 郑新建.一种基于流水线结构的多级数字混沌编码方案[J].计算机技术与发展,2007,(05):152.
　ZHANG Xiao-qi,ZHANG Yi-wei,ZHENG Xin-jian.A Multilevel Digital Chaotic Encoding Scheme Based on Pipeline Structure[J].,2007,(06):152.
[9]牛志伟 黄红女.Windows平台下机群并行编译环境配置[J].计算机技术与发展,2007,(08):15.
　NIU Zhi-wei,HUANG Hong-nü.Configuration of Parallel Compile Environment of Cluster on Windows Platform[J].,2007,(06):15.
[10]蔡佳佳 李名世 郑锋.多核微机基于OpenMP的并行计算[J].计算机技术与发展,2007,(10):87.
　CAI Jia-jia,LI Ming-shi,ZHENG Feng.OpenMP- Based Parallel Computation on Multi- Core PC[J].,2007,(06):87.
[11]胥秀峰,鲍广宇,黄海燕,等. 基于CMP系统的并行编程模式研究[J].计算机技术与发展,2014,24(07):80.
　XU Xiu-feng,BAO Guang-yu,HUANG Hai-yan,et al. Study on Patterns for Parallel Programming Based on CMP System[J].,2014,24(06):80.
[12]王俊超,彭涛,冯光柳. 曙光高性能计算机在数值预报模式中的应用[J].计算机技术与发展,2014,24(10):178.
　WANG Jun-chao,PENG Tao,FENG Guang-liu. Application of HPC in Numerical Prediction Model[J].,2014,24(06):178.
[13]陈树敏[][],罗俊博[] [],陈青[]. 并行计算技术的几种实现方式研究[J].计算机技术与发展,2015,25(09):174.
　CHEN Shu-min[] [],LUO Jun-bo[] [],CHEN Qing[]. Research on Ways of Parallel Computing Technology[J].,2015,25(06):174.
[14]陈静,郑彦. 基于二叉树的并行频繁项集挖掘算法[J].计算机技术与发展,2015,25(10):80.
　CHEN Jing,ZHENG Yan. Parallel Algorithm of Frequent Itemset Mining Based on Binary-tree[J].,2015,25(06):80.
[15]徐辉,王宁章,雷琳琳. 一种海量中文地址转化与切割的方法研究[J].计算机技术与发展,2015,25(11):6.
　XU Hui,WANG Ning-zhang,LEI Lin-lin. Research on a Massive Chinese Address Conversion and Cutting Method[J].,2015,25(06):6.
[16]孙建[],李涛[],李雪丹[]. 基于PAAG的图形图像算法的并行实现[J].计算机技术与发展,2015,25(11):61.
　SUN Jian[],LI Tao[],LI Xue-dan[]. Parallel Implementation of Graphics Rendering and Image Processing Algorithm Based on PAAG[J].,2015,25(06):61.
[17]刘永江[],邵庆[],彭淑罗[]. 分布式地震数据文件系统[J].计算机技术与发展,2015,25(11):209.
　LIU Yong-jiang[],SHAO Qing[],PENG Shu-luo[]. Distributed Seismic Data File System[J].,2015,25(06):209.
[18]李雪迪,郑彦. 基于分布式倒排索引的频繁项集挖掘[J].计算机技术与发展,2016,26(03):101.
　LI Xue-di,ZHENG Yan. Frequent Itemset Mining Based on Distributed Inverted Index[J].,2016,26(06):101.

更新日期/Last Update: 2017-07-26