Simulation Based Performance Analysis of Ethernet MPI Cluster

Boleslaw Tokarski, L. Koszalka
{"title":"Simulation Based Performance Analysis of Ethernet MPI Cluster","authors":"Boleslaw Tokarski, L. Koszalka","doi":"10.1109/ICN.2009.55","DOIUrl":null,"url":null,"abstract":"This paper shows the influence of network aspects and MPI implementation on the performance of an Ethernet based computer cluster. Following factors are considered: message size, number of nodes, node heterogeneity, network speed,switching technology and MPI implementation. The basic index of performance is throughput measured with Intel MPI Benchmark.This paper shows that there is a specific message size that is optimal for each cluster configuration. Most important factor is the network speed, the second is MPI implementation. One can also find that using a faster master node can give a speed bonus.","PeriodicalId":299215,"journal":{"name":"2009 Eighth International Conference on Networks","volume":"163 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2009-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2009 Eighth International Conference on Networks","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICN.2009.55","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

This paper shows the influence of network aspects and MPI implementation on the performance of an Ethernet based computer cluster. Following factors are considered: message size, number of nodes, node heterogeneity, network speed,switching technology and MPI implementation. The basic index of performance is throughput measured with Intel MPI Benchmark.This paper shows that there is a specific message size that is optimal for each cluster configuration. Most important factor is the network speed, the second is MPI implementation. One can also find that using a faster master node can give a speed bonus.
基于仿真的以太网MPI集群性能分析
本文介绍了网络方面和MPI实现对基于以太网的计算机集群性能的影响。考虑以下因素:消息大小、节点数量、节点异构性、网络速度、交换技术和MPI实现。性能的基本指标是用Intel MPI Benchmark测量的吞吐量。本文表明,对于每个集群配置,有一个特定的消息大小是最优的。最重要的因素是网络速度,其次是MPI的实现。我们还可以发现,使用更快的主节点可以获得速度奖励。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信