Design and Performance Analysis of Low Latency Routing Algorithm based NoC for MPSoC

T. Nagalaxmi, E. S. Rao, P. Chandrasekhar
{"title":"Design and Performance Analysis of Low Latency Routing Algorithm based NoC for MPSoC","authors":"T. Nagalaxmi, E. S. Rao, P. Chandrasekhar","doi":"10.17762/ijcnis.v14i1s.5590","DOIUrl":null,"url":null,"abstract":"The Network on Chip is appropriate where System-on-Chip technology is scalable and adaptable. The Network on Chip is a new communication architecture with a number of benefits, including scalability, flexibility, and reusability, for applications built on Multiprocessor System on a Chip (MPSoC). However, the design of efficient NoC fabric with high performance is critically complex because of its architectural parameters. Identifying a suitable scheduling algorithm to resolve arbitration among ports to obtain high-speed data transfer in the router is one of the most significant phases while designing a Network on chip based Multiprocessor System on a Chip. Low latency, throughput, space utilization, energy consumption, and reliability for Network on chip fabric are all determined by the router. The performance of the NoC system is hampered by the deadlock issues that plague conventional routing algorithms. This work develops a novel routing algorithm to address the deadlock problem. In this paper, a deterministic shortest path deadlock-free routing method is developed based on the analysis of the Turn Model. In the 2D-mesh structure, the algorithm uses separate routing methods for the odd and even columns. This minimizes the number of paths for a single channel, congestion, and latency. Two test scenarios—one with and one without a load test—were used to evaluate the proposed model. For a zero-load network, three clock cycles are utilized to transfer the packets. For the load network, five clocks are utilized to transfer the packets. The latency is measured for both cases without load and with load test and the corresponding latency is 3ns and 7ns respectively.The proposed method has an 18.57Mbps throughput.  The area and power utilization for the proposed method are 69% (IO utilization) and 0.128W respectively. In order to validate the proposed method, the latency is compared with existing work and 50% latency is reduced both with and without congestion load.","PeriodicalId":232613,"journal":{"name":"Int. J. Commun. Networks Inf. Secur.","volume":"81 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Int. J. Commun. Networks Inf. Secur.","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.17762/ijcnis.v14i1s.5590","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

The Network on Chip is appropriate where System-on-Chip technology is scalable and adaptable. The Network on Chip is a new communication architecture with a number of benefits, including scalability, flexibility, and reusability, for applications built on Multiprocessor System on a Chip (MPSoC). However, the design of efficient NoC fabric with high performance is critically complex because of its architectural parameters. Identifying a suitable scheduling algorithm to resolve arbitration among ports to obtain high-speed data transfer in the router is one of the most significant phases while designing a Network on chip based Multiprocessor System on a Chip. Low latency, throughput, space utilization, energy consumption, and reliability for Network on chip fabric are all determined by the router. The performance of the NoC system is hampered by the deadlock issues that plague conventional routing algorithms. This work develops a novel routing algorithm to address the deadlock problem. In this paper, a deterministic shortest path deadlock-free routing method is developed based on the analysis of the Turn Model. In the 2D-mesh structure, the algorithm uses separate routing methods for the odd and even columns. This minimizes the number of paths for a single channel, congestion, and latency. Two test scenarios—one with and one without a load test—were used to evaluate the proposed model. For a zero-load network, three clock cycles are utilized to transfer the packets. For the load network, five clocks are utilized to transfer the packets. The latency is measured for both cases without load and with load test and the corresponding latency is 3ns and 7ns respectively.The proposed method has an 18.57Mbps throughput.  The area and power utilization for the proposed method are 69% (IO utilization) and 0.128W respectively. In order to validate the proposed method, the latency is compared with existing work and 50% latency is reduced both with and without congestion load.
基于NoC的MPSoC低延迟路由算法设计与性能分析
片上网络适用于片上系统技术具有可扩展性和适应性的地方。片上网络是一种新的通信架构,具有许多优点,包括可扩展性、灵活性和可重用性,适用于构建在多处理器片上系统(MPSoC)上的应用程序。然而,由于其结构参数的限制,高效高性能NoC结构的设计非常复杂。在设计基于片上多处理器系统的片上网络时,确定合适的调度算法来解决路由器中端口间的仲裁,以获得高速数据传输是最重要的阶段之一。片上网络结构的低延迟、吞吐量、空间利用率、能耗和可靠性都由路由器决定。NoC系统的性能受到困扰传统路由算法的死锁问题的影响。本文提出了一种新的路由算法来解决死锁问题。本文在分析转弯模型的基础上,提出了一种确定性最短路径无死锁路由方法。在二维网格结构中,该算法对奇数列和偶数列采用单独的路由方法。这样可以最大限度地减少单个通道的路径数量、拥塞和延迟。两个测试场景——一个有负载测试,一个没有负载测试——被用来评估提议的模型。对于零负载网络,使用3个时钟周期来传输数据包。对于负载网络,使用五个时钟来传输数据包。在无负载和有负载测试两种情况下,延迟分别为3ns和7ns。该方法的吞吐量为18.57Mbps。该方法的面积和功率利用率分别为69% (IO利用率)和0.128W。为了验证所提出的方法,将延迟与现有工作进行了比较,在有和没有拥塞负载的情况下,延迟都减少了50%。
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