On writing a router for message passing in a transputer network

A.Augusto Sousa, F.Nunes Ferreira
{"title":"On writing a router for message passing in a transputer network","authors":"A.Augusto Sousa,&nbsp;F.Nunes Ferreira","doi":"10.1016/0956-0521(95)00038-0","DOIUrl":null,"url":null,"abstract":"<div><p>In a distributed memory MIMD parallel machine, the efficient communication between processes/processors, through messages, is an important task to be handled by the programmer. Because the number of inter-processor connections is limited, the communication between any two processors is made by passing the messages through several other processors and then, a problem of messages routing appears. For dedicated systems, special architectures can be defined simplifying the problem but, if an environment constituting a basis for general applications development is desired, the problem is more serious due to the deadlock possibility. A general router, able to avoid the problem, becomes then a very important tool for software development in parallel architectures. We have been defining a development platform, based on a network of Transputers and written in OCCAM, for image synthesis applications. This paper reports our efforts in writing different versions of routers, based on two different strategies, and justifies the choice of an efficient one to integrate in the platform.</p></div>","PeriodicalId":100325,"journal":{"name":"Computing Systems in Engineering","volume":"6 4","pages":"Pages 471-476"},"PeriodicalIF":0.0000,"publicationDate":"1995-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1016/0956-0521(95)00038-0","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computing Systems in Engineering","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/0956052195000380","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1

Abstract

In a distributed memory MIMD parallel machine, the efficient communication between processes/processors, through messages, is an important task to be handled by the programmer. Because the number of inter-processor connections is limited, the communication between any two processors is made by passing the messages through several other processors and then, a problem of messages routing appears. For dedicated systems, special architectures can be defined simplifying the problem but, if an environment constituting a basis for general applications development is desired, the problem is more serious due to the deadlock possibility. A general router, able to avoid the problem, becomes then a very important tool for software development in parallel architectures. We have been defining a development platform, based on a network of Transputers and written in OCCAM, for image synthesis applications. This paper reports our efforts in writing different versions of routers, based on two different strategies, and justifies the choice of an efficient one to integrate in the platform.

关于编写一个路由器用于在计算机网络中传递消息
在分布式内存MIMD并行机中,进程/处理器之间通过消息的有效通信是程序员需要处理的一项重要任务。由于处理器间连接的数量有限,任何两个处理器之间的通信都是通过将消息传递给其他几个处理器来实现的,这样就出现了消息路由问题。对于专用系统,可以定义特殊的体系结构来简化问题,但是,如果需要一个构成一般应用程序开发基础的环境,则由于死锁的可能性,问题会更加严重。一个通用的路由器,能够避免这个问题,因此成为并行架构下软件开发的一个非常重要的工具。我们一直在定义一个基于Transputers网络并使用OCCAM编写的用于图像合成应用程序的开发平台。本文报告了我们基于两种不同策略编写不同版本路由器的努力,并证明了选择一种有效的路由器集成到平台中。
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:481959085
Book学术官方微信