Introduction to the special issue on application-specific processors

P. Brisk, T. Mitra
{"title":"Introduction to the special issue on application-specific processors","authors":"P. Brisk, T. Mitra","doi":"10.1145/2514641.2514642","DOIUrl":null,"url":null,"abstract":"Introduction to the Special Issue on Application-Specific Processors Application-specific processors offer performance, energy, and cost benefits compared to their general-purpose counterparts for a wide variety of market segments, ranging from low-cost micro-controllers to high-end supercomputers. The design and usage of application-specific processors are strikingly different from general-purpose comput- ers as the architecture is tuned to accelerate only a specific application or a class of applications. Application-specific architectures are often derived from a high-level specification of an application through a hardware/software co-design process, which can be highly automated; such a process would be inappropriate for general-purpose processor de- sign. Moreover, code compilation targeting application-specific processors is often inter- twined with the co-design process. Lastly, the system software may provide reduced and targeted functionality and interfaces at a much lower cost than a traditional embedded or real-time operating system such as iOS or Android. Historically, application-specific processors have been part of low-cost embedded devices; however, the landscape is rapidly changing. One important factor is the emer- gence of graphics processing units (GPUs) as general-purpose high-performance computing devices, which utilize graphics-specific hardware, but can also provide sig- nificant acceleration boosts for vectorizable applications. Similarly, reconfigurable computing devices, such as FPGAs, provide a flexible hardware fabric that can be used to implement a wide variety of application-specific functionality and as such are increasingly used for code acceleration in different domains. This special issue of ACM Transactions on Embedded Computing Systems is dedi- cated to all aspects of application-specific processors. Part of this special issue presents extended versions of some of the best papers that were presented at the IEEE Sym- posium on Application-Specific Processors (SASP) in 2009 and 2010. Altogether, we received 66 submitted manuscripts for the special issue, 10 of which were accepted for inclusion in this special issue. It is our great honor to introduce the articles. Our first article, “Hardware Architectural Support for Control Systems and Sensor Processing” by Sundhanshu Vyas, Adwait Gupta, Christopher Gill, Ron K. Cytron, Joseph Zambreno and Philip Jones, describes a microcontroller that has been cus- tomized to control thousands of PID controllers concurrently. Most application-specific processors in the past have extended RISCs or VLIWs. This article represents a land- mark effort in the design of customizable application-specific microcontrollers. This special issue features three articles that focus on the architecture of application- specific processors and the design of application-specific accelerators, namely, “Multicore-Based Vector Coprocessor Sharing for Performance and Energy Gains” by Spiridon F. Beldianu and Sotirios G. Ziavras, “A Systematic Approach for Optimized By- pass Configurations for Application-Specific Embedded Processors” by Thorsten Junge- blut, Boris Hubener, Mario Porrmann, and Ulrich R uckert, and “Custom Architecture for Multicore Audio Beamforming Systems” by Dimitris Theodoropoulos and Georgi Kuzmanov. They all represent interesting design points for different application- specific architectural styles. This is a large design space, and these three articles provide excellent insights about how one can customize a hardware platform. c 2013 ACM 1539-9087/2013/09-ART15 $15.00 DOI: http://dx.doi.org/10.1145/2514641.2514642 ACM Transactions on Embedded Computing Systems, Vol. 13, No. 2, Article 15, Publication date: September 2013.","PeriodicalId":183677,"journal":{"name":"ACM Trans. Embed. Comput. Syst.","volume":"10 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2013-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"ACM Trans. Embed. Comput. Syst.","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/2514641.2514642","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

Introduction to the Special Issue on Application-Specific Processors Application-specific processors offer performance, energy, and cost benefits compared to their general-purpose counterparts for a wide variety of market segments, ranging from low-cost micro-controllers to high-end supercomputers. The design and usage of application-specific processors are strikingly different from general-purpose comput- ers as the architecture is tuned to accelerate only a specific application or a class of applications. Application-specific architectures are often derived from a high-level specification of an application through a hardware/software co-design process, which can be highly automated; such a process would be inappropriate for general-purpose processor de- sign. Moreover, code compilation targeting application-specific processors is often inter- twined with the co-design process. Lastly, the system software may provide reduced and targeted functionality and interfaces at a much lower cost than a traditional embedded or real-time operating system such as iOS or Android. Historically, application-specific processors have been part of low-cost embedded devices; however, the landscape is rapidly changing. One important factor is the emer- gence of graphics processing units (GPUs) as general-purpose high-performance computing devices, which utilize graphics-specific hardware, but can also provide sig- nificant acceleration boosts for vectorizable applications. Similarly, reconfigurable computing devices, such as FPGAs, provide a flexible hardware fabric that can be used to implement a wide variety of application-specific functionality and as such are increasingly used for code acceleration in different domains. This special issue of ACM Transactions on Embedded Computing Systems is dedi- cated to all aspects of application-specific processors. Part of this special issue presents extended versions of some of the best papers that were presented at the IEEE Sym- posium on Application-Specific Processors (SASP) in 2009 and 2010. Altogether, we received 66 submitted manuscripts for the special issue, 10 of which were accepted for inclusion in this special issue. It is our great honor to introduce the articles. Our first article, “Hardware Architectural Support for Control Systems and Sensor Processing” by Sundhanshu Vyas, Adwait Gupta, Christopher Gill, Ron K. Cytron, Joseph Zambreno and Philip Jones, describes a microcontroller that has been cus- tomized to control thousands of PID controllers concurrently. Most application-specific processors in the past have extended RISCs or VLIWs. This article represents a land- mark effort in the design of customizable application-specific microcontrollers. This special issue features three articles that focus on the architecture of application- specific processors and the design of application-specific accelerators, namely, “Multicore-Based Vector Coprocessor Sharing for Performance and Energy Gains” by Spiridon F. Beldianu and Sotirios G. Ziavras, “A Systematic Approach for Optimized By- pass Configurations for Application-Specific Embedded Processors” by Thorsten Junge- blut, Boris Hubener, Mario Porrmann, and Ulrich R uckert, and “Custom Architecture for Multicore Audio Beamforming Systems” by Dimitris Theodoropoulos and Georgi Kuzmanov. They all represent interesting design points for different application- specific architectural styles. This is a large design space, and these three articles provide excellent insights about how one can customize a hardware platform. c 2013 ACM 1539-9087/2013/09-ART15 $15.00 DOI: http://dx.doi.org/10.1145/2514641.2514642 ACM Transactions on Embedded Computing Systems, Vol. 13, No. 2, Article 15, Publication date: September 2013.
介绍特定于应用程序的处理器的特殊问题
从低成本微控制器到高端超级计算机,与通用处理器相比,专用处理器在性能、能源和成本方面都有优势。应用程序专用处理器的设计和使用与通用计算机有着显著的不同,因为其体系结构被调整为只加速特定的应用程序或一类应用程序。特定于应用程序的体系结构通常是通过硬件/软件协同设计过程从应用程序的高级规范派生出来的,该过程可以高度自动化;这样的过程不适合通用处理器的设计。此外,针对特定应用程序处理器的代码编译通常与协同设计过程交织在一起。最后,系统软件可以以比传统的嵌入式或实时操作系统(如iOS或Android)低得多的成本提供更少的目标功能和接口。从历史上看,专用处理器一直是低成本嵌入式设备的一部分;然而,形势正在迅速变化。一个重要的因素是图形处理单元(gpu)作为通用高性能计算设备的出现,它利用特定于图形的硬件,但也可以为可矢量化的应用程序提供显著的加速提升。类似地,可重构计算设备(如fpga)提供了灵活的硬件结构,可用于实现各种特定于应用程序的功能,因此越来越多地用于不同领域的代码加速。这一期ACM嵌入式计算系统事务专刊专门介绍了应用程序专用处理器的各个方面。本期特刊的部分内容是2009年和2010年IEEE专用处理器研讨会(SASP)上发表的一些最佳论文的扩展版本。本次特刊共收到投稿稿件66篇,其中10篇入选。我们很荣幸向您介绍这些商品。我们的第一篇文章,“控制系统和传感器处理的硬件架构支持”,由Sundhanshu Vyas, Adwait Gupta, Christopher Gill, Ron K. Cytron, Joseph Zambreno和Philip Jones撰写,描述了一个微控制器,该微控制器已被定制为同时控制数千个PID控制器。过去,大多数特定于应用程序的处理器都扩展了risc或vliw。这篇文章代表了在设计可定制的特定应用微控制器方面的一个里程碑式的努力。这期特刊有三篇文章,重点关注特定应用处理器的架构和特定应用加速器的设计,即Spiridon F. Beldianu和Sotirios G. Ziavras的“基于多核的矢量协同处理器共享的性能和能量增益”,由Thorsten Junge- blut, Boris Hubener, Mario Porrmann和Ulrich R uckert撰写的“特定应用嵌入式处理器优化旁路配置的系统方法”,以及Dimitris Theodoropoulos和Georgi Kuzmanov的“多核音频波束形成系统的自定义架构”。它们都代表了不同应用程序特定架构风格的有趣设计点。这是一个很大的设计空间,这三篇文章提供了关于如何定制硬件平台的极好见解。c 2013 ACM 1539-9087/2013/09-ART15 $15.00 DOI: http://dx.doi.org/10.1145/2514641.2514642 ACM Transactions on Embedded Computing Systems, Vol. 13, No. 2, Article 15,出版日期:2013年9月。
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