[1]崔奇,谷建华.MPI集合通信剖析技术的研究[J].计算机技术与发展,2013,(10):31-35.
　CUI Qi,GU Jian-hua.Research of Collective Communications Profiling in MPI[J].,2013,(10):31-35.
点击复制

MPI集合通信剖析技术的研究()
分享到：

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

卷:
期数:
2013年10期
页码:
31-35
栏目:
智能、算法、系统工程
出版日期:
1900-01-01

文章信息/Info

Title:
Research of Collective Communications Profiling in MPI
文章编号:
1673-629X（2013）10-0031-05
作者:
崔奇; 谷建华
西北工业大学 计算机学院
Author(s):
CUI Qi; GU Jian-hua
关键词:
MPI; 集合通信; 通信剖析; 进程映射
Keywords:
MPI; collective communication; communication profiling; process mapping
文献标志码:
A
摘要:
将MPI( Message Passing Interface)进程拓扑有效地映射到处理器拓扑上有助于提高MPI程序的通信性能。目前大部分的MPI进程映射只考虑点对点通信，很少考虑到集合通信，原因是获取集合通信的进程拓扑是比较困难的。目前大部分剖析( profiling)工具在剖析集合通信时只考虑了函数的接口语义，而忽视了实现语义，导致这些工具不能正确地获取集合通信进程之间的详细通信情况。文中提出了一套剖析算法，可以准确地计算出参与集合通信的每对进程之间的通信量，并以通信矩阵的形式给出进程拓扑。实验证明了剖析算法的正确性，并且通过这种剖析方法获取的进程拓扑能够提升进程到处理器核的映射实验效果
Abstract:
The mapping from processes to processors can help the Message Passing Interface ( MPI) application improve the communica-tion performance of parallel program. Most of the mappings at present are based on point-to-point communications,and barely consider the collective communications,because it is difficult to obtain the process topology of collective communications. Most of the profiling tools nowadays only think over the interface semantics of functions in analyzing collective communications,and ignore the implementation semantics,leading to these tools can't acquire detailed communication situation of processes accurately in collective communications. Present a set of profiling algorithm which can obtain the communication traffic for each pair of processes participating collective commu-nications accurately. The experimental results show the correctness of the profiling method,and the process topology acquired through this profiling method can get better effect in the mapping from process to core

相似文献/References:

[1]卢照 师军 于海蛟 方昕.城市路网的最短路径并行求解[J].计算机技术与发展,2010,(01):78.
　LU Zhao,SHI Jun,YU Hai-jiao,et al.Solving the Shortest Path in Parallel of City Road Network[J].,2010,(10):78.
[2]胡斌 袁道华.TBB多核编程及其混合编程模型的研究[J].计算机技术与发展,2009,(02):98.
　HU Bin,YUAN Dao-hua.TBB Multi- Core Progamming and Research on Its Hybrid Paradigms[J].,2009,(10):98.
[3]胡晨骏 王晓蔚.基于多核集群系统的并行编程模型的研究[J].计算机技术与发展,2008,(04):70.
　HU Chen-jun,WANG Xiao-wei.Research of Parallel Programming Model Based on Multi-Core Cluster System[J].,2008,(10):70.
[4]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,(10):15.
[5]宋伟 宋玉.基于SMP集群系统的并行编程模式研究与分析[J].计算机技术与发展,2007,(02):164.
　SONG Wei,SONG Yu.Research and Analysis of Parallel Programming Model Based on SMP Cluster System[J].,2007,(10):164.
[6]黎凯伦 吴伟民 苏庆.按行及按列划分的并行矩阵向量乘法的分析[J].计算机技术与发展,2006,(04):41.
　LI Kai-lun,WU Wei-min,SU Qing.Analysis on Parallel Matrix- Vector Multiplication Based on Divided by Row and Divided by Column[J].,2006,(10):41.
[7]张翠莲 刘方爱 王亚楠.基于MPI的并行程序设计[J].计算机技术与发展,2006,(08):72.
　ZHANG Cui-lian,LIU Fang-ai,WANG Ya-nan.Parallel Program Design Based on MPI[J].,2006,(10):72.
[8]李海荣,方中纯.应用MPICH并行化图像处理算法[J].计算机技术与发展,2013,(07):101.
　LI Hai-rong,FANG Zhong-chun.Parallelizing Algorithms of Image Processing Using MPICH[J].,2013,(10):101.

更新日期/Last Update: 1900-01-01