Ivan Luiz Pedroso Pires, Marco Antonio Zanata Alves, Luiz Carlos Pessoa Albini
{"title":"Trace-driven and processing time extensions for Noxim simulator","authors":"Ivan Luiz Pedroso Pires, Marco Antonio Zanata Alves, Luiz Carlos Pessoa Albini","doi":"10.1007/s10617-018-09218-7","DOIUrl":null,"url":null,"abstract":"Simulation is one of the main tools used to analyze and test new proposals in the Network-on-Chip field. Several simulators can be found in the literature, among them the Noxim simulator stands out. It is being used by many researchers due to the wireless support and open-source availability. An important issue at the simulation phase is the choice of workload, as it may affect testing the system and its features. The correct workload can lead to rapid and efficient system development, while the wrong one may compromise the entire system evaluation. To ensure a more realistic simulation, simulators usually relies on real workloads by using a trace-driven approach. Although Noxim provides a simple support for input traces, it is very limited to a general behavior of the system, accepting only a generic injection rate parameter over time. Another important part of the simulator is the ability to consider the Processing Elements processing time. We propose in this paper an extension of the Noxim simulator to address these issues. Consequently, results are more realistic and may be possible to predict the total execution time very accurately. This extension is demonstrated and evaluated using the NAS-NPB workload.","PeriodicalId":50594,"journal":{"name":"Design Automation for Embedded Systems","volume":"7 4","pages":"41-55"},"PeriodicalIF":0.9000,"publicationDate":"2019-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Design Automation for Embedded Systems","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.1007/s10617-018-09218-7","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"COMPUTER SCIENCE, HARDWARE & ARCHITECTURE","Score":null,"Total":0}
引用次数: 2
Abstract
Simulation is one of the main tools used to analyze and test new proposals in the Network-on-Chip field. Several simulators can be found in the literature, among them the Noxim simulator stands out. It is being used by many researchers due to the wireless support and open-source availability. An important issue at the simulation phase is the choice of workload, as it may affect testing the system and its features. The correct workload can lead to rapid and efficient system development, while the wrong one may compromise the entire system evaluation. To ensure a more realistic simulation, simulators usually relies on real workloads by using a trace-driven approach. Although Noxim provides a simple support for input traces, it is very limited to a general behavior of the system, accepting only a generic injection rate parameter over time. Another important part of the simulator is the ability to consider the Processing Elements processing time. We propose in this paper an extension of the Noxim simulator to address these issues. Consequently, results are more realistic and may be possible to predict the total execution time very accurately. This extension is demonstrated and evaluated using the NAS-NPB workload.
期刊介绍：
Embedded (electronic) systems have become the electronic engines of modern consumer and industrial devices, from automobiles to satellites, from washing machines to high-definition TVs, and from cellular phones to complete base stations. These embedded systems encompass a variety of hardware and software components which implement a wide range of functions including digital, analog and RF parts.
Although embedded systems have been designed for decades, the systematic design of such systems with well defined methodologies, automation tools and technologies has gained attention primarily in the last decade. Advances in silicon technology and increasingly demanding applications have significantly expanded the scope and complexity of embedded systems. These systems are only now becoming possible due to advances in methodologies, tools, architectures and design techniques.
Design Automation for Embedded Systems is a multidisciplinary journal which addresses the systematic design of embedded systems, focusing primarily on tools, methodologies and architectures for embedded systems, including HW/SW co-design, simulation and modeling approaches, synthesis techniques, architectures and design exploration, among others.
Design Automation for Embedded Systems offers a forum for scientist and engineers to report on their latest works on algorithms, tools, architectures, case studies and real design examples related to embedded systems hardware and software.
Design Automation for Embedded Systems is an innovative journal which distinguishes itself by welcoming high-quality papers on the methodology, tools, architectures and design of electronic embedded systems, leading to a true multidisciplinary system design journal.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
