使用虚拟样机(ViPro)工具设计寄存器文件吞吐量和能量的优化

Ningxi Liu, B. Calhoun
{"title":"使用虚拟样机(ViPro)工具设计寄存器文件吞吐量和能量的优化","authors":"Ningxi Liu, B. Calhoun","doi":"10.1109/ISVLSI.2016.50","DOIUrl":null,"url":null,"abstract":"Register files (RFs) consume significant power in low-power processors, and their specifications vary substantially for different applications. Challenges exist in identifying the appropriate RF design and optimizing RFs for different specifications. This paper not only explores methodologies of designing low power and high performance RFs and it also extends a virtual prototyping (ViPro) tool to support fast and efficient estimation of different design knobs on the overall multi-port RF macros. To enable aggressive exploration for RFs design, three bitline (BL) sensing schemes are included into ViPro along with parasitic parameters extracted from layout. Accuracy of ViPro results are within 15 % compared to full RF schematic SPICE simulation, while the simulation speed of ViPro is 5-10 times faster. An example reveals how ViPro can optimize RF design based on various specifications in a 45nm CMOS technology. Improvements of data throughput for 1R/1W port RFs are 31% and 72% at 0.5KB and 512KB, respectively, with proper BL sensing techniques. Results also show that the optimal BL sensing scheme changes with memory capacity. At 0.5KB, the lowest energy per operation decreases by 7.5% with a single-ended BL, while energy reduction is 45% with a hierarchical BL for 512KB.","PeriodicalId":140647,"journal":{"name":"2016 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2016-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"Design Optimization of Register File Throughput and Energy Using a Virtual Prototyping (ViPro) Tool\",\"authors\":\"Ningxi Liu, B. Calhoun\",\"doi\":\"10.1109/ISVLSI.2016.50\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Register files (RFs) consume significant power in low-power processors, and their specifications vary substantially for different applications. Challenges exist in identifying the appropriate RF design and optimizing RFs for different specifications. This paper not only explores methodologies of designing low power and high performance RFs and it also extends a virtual prototyping (ViPro) tool to support fast and efficient estimation of different design knobs on the overall multi-port RF macros. To enable aggressive exploration for RFs design, three bitline (BL) sensing schemes are included into ViPro along with parasitic parameters extracted from layout. Accuracy of ViPro results are within 15 % compared to full RF schematic SPICE simulation, while the simulation speed of ViPro is 5-10 times faster. An example reveals how ViPro can optimize RF design based on various specifications in a 45nm CMOS technology. Improvements of data throughput for 1R/1W port RFs are 31% and 72% at 0.5KB and 512KB, respectively, with proper BL sensing techniques. Results also show that the optimal BL sensing scheme changes with memory capacity. At 0.5KB, the lowest energy per operation decreases by 7.5% with a single-ended BL, while energy reduction is 45% with a hierarchical BL for 512KB.\",\"PeriodicalId\":140647,\"journal\":{\"name\":\"2016 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)\",\"volume\":\"27 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2016-07-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2016 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ISVLSI.2016.50\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2016 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISVLSI.2016.50","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1

摘要

寄存器文件(RFs)在低功耗处理器中消耗大量的功率,并且它们的规格在不同的应用程序中有很大的不同。在确定合适的射频设计和优化不同规格的射频方面存在挑战。本文不仅探讨了设计低功耗和高性能RF的方法,而且还扩展了虚拟样机(ViPro)工具,以支持对整体多端口RF宏的不同设计旋钮进行快速有效的估计。为了能够积极探索rf设计,ViPro中包含了三种位线(BL)传感方案以及从布局中提取的寄生参数。与全RF原理图SPICE模拟相比,ViPro结果的准确性在15%以内,而ViPro的模拟速度快5-10倍。一个例子揭示了ViPro如何在45nm CMOS技术中基于各种规格优化RF设计。采用适当的BL传感技术,1R/1W端口rf在0.5KB和512KB时的数据吞吐量分别提高了31%和72%。结果还表明,最优BL感知方案随存储容量的变化而变化。在0.5KB时,单端BL每次操作的最低能量减少了7.5%,而在512KB时,分层BL的能量减少了45%。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Design Optimization of Register File Throughput and Energy Using a Virtual Prototyping (ViPro) Tool
Register files (RFs) consume significant power in low-power processors, and their specifications vary substantially for different applications. Challenges exist in identifying the appropriate RF design and optimizing RFs for different specifications. This paper not only explores methodologies of designing low power and high performance RFs and it also extends a virtual prototyping (ViPro) tool to support fast and efficient estimation of different design knobs on the overall multi-port RF macros. To enable aggressive exploration for RFs design, three bitline (BL) sensing schemes are included into ViPro along with parasitic parameters extracted from layout. Accuracy of ViPro results are within 15 % compared to full RF schematic SPICE simulation, while the simulation speed of ViPro is 5-10 times faster. An example reveals how ViPro can optimize RF design based on various specifications in a 45nm CMOS technology. Improvements of data throughput for 1R/1W port RFs are 31% and 72% at 0.5KB and 512KB, respectively, with proper BL sensing techniques. Results also show that the optimal BL sensing scheme changes with memory capacity. At 0.5KB, the lowest energy per operation decreases by 7.5% with a single-ended BL, while energy reduction is 45% with a hierarchical BL for 512KB.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
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:604180095
Book学术官方微信