The polybus: A flexible and fault-tolerant multiprocessor interconnection

R Männer, B Deluigi, W Saaler, T Sauer, P.V Walter
{"title":"The polybus: A flexible and fault-tolerant multiprocessor interconnection","authors":"R Männer,&nbsp;B Deluigi,&nbsp;W Saaler,&nbsp;T Sauer,&nbsp;P.V Walter","doi":"10.1016/0252-7308(84)90034-5","DOIUrl":null,"url":null,"abstract":"<div><p>Polybus is a general purpose multiprocessor bus system which links modules such as processors, memories and input-output devices in a homogeneous way. It is a multiple common bus consisting of an arbitrary number of busses capable of independent operation. Connected modules, however, access the Polybus system like a single common bus. The multiple bus structure is user transparent, <em>i.e</em> busses may be added or removed without hardware or software modifications. The bus bandwidth can therefore be adjusted to actual requirements. Transfer speeds of up to <em>n</em> × 12 Mbytes s<sup>−1</sup> (where <em>n</em> is the number of busses used) can be achieved. Circuit switching through the bus system is managed by a decentralized arbiter which allows access according to priority as well as fair treatment of requests. Random assignment of busses to requests enables automatic by passing of faulty paths by re-try. Several addressing modes allow access to specific modules, to pools of modules or to an arbitrary member of a pool. This supports multiprocessor-transparent programming as well as data flow applications. In this paper an overview of the Polybus system, details of its operation and a comparison of some implementation details with other standard solutions are presented.</p></div>","PeriodicalId":100687,"journal":{"name":"Interfaces in Computing","volume":"2 1","pages":"Pages 45-68"},"PeriodicalIF":0.0000,"publicationDate":"1984-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1016/0252-7308(84)90034-5","citationCount":"14","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Interfaces in Computing","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/0252730884900345","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 14

Abstract

Polybus is a general purpose multiprocessor bus system which links modules such as processors, memories and input-output devices in a homogeneous way. It is a multiple common bus consisting of an arbitrary number of busses capable of independent operation. Connected modules, however, access the Polybus system like a single common bus. The multiple bus structure is user transparent, i.e busses may be added or removed without hardware or software modifications. The bus bandwidth can therefore be adjusted to actual requirements. Transfer speeds of up to n × 12 Mbytes s−1 (where n is the number of busses used) can be achieved. Circuit switching through the bus system is managed by a decentralized arbiter which allows access according to priority as well as fair treatment of requests. Random assignment of busses to requests enables automatic by passing of faulty paths by re-try. Several addressing modes allow access to specific modules, to pools of modules or to an arbitrary member of a pool. This supports multiprocessor-transparent programming as well as data flow applications. In this paper an overview of the Polybus system, details of its operation and a comparison of some implementation details with other standard solutions are presented.

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
多总线:一种灵活的、容错的多处理器互连
Polybus是一种通用的多处理器总线系统,它以一种同质的方式连接处理器、存储器和输入输出设备等模块。它是由任意数量的能够独立运行的总线组成的多总线。然而,连接的模块像一个单一的公共总线一样访问Polybus系统。多总线结构是用户透明的,即可以在不修改硬件或软件的情况下增加或删除总线。因此,总线带宽可以根据实际需要进行调整。传输速率最高可达n × 12mb s−1 (n为使用的总线数)。通过总线系统的电路交换由一个分散的仲裁器管理,该仲裁器允许根据优先级访问以及公平处理请求。随机分配总线到请求,通过重试通过故障路径实现自动。有几种寻址模式允许访问特定模块、模块池或池中的任意成员。这支持多处理器透明编程以及数据流应用程序。本文概述了Polybus系统,详细介绍了它的操作,并将一些实现细节与其他标准解决方案进行了比较。
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 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
Subject index Author index Machine-independent interfaces to networking software Networked IBM 3270 emulation on a microcomputer Command line structure and dynamic processing of abbreviations in dialogue management
×
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
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 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 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1