A hardware acceleration technique for gradient descent and conjugate gradient

David Kesler, Biplab Deka, Rakesh Kumar
{"title":"A hardware acceleration technique for gradient descent and conjugate gradient","authors":"David Kesler, Biplab Deka, Rakesh Kumar","doi":"10.1109/SASP.2011.5941086","DOIUrl":null,"url":null,"abstract":"Application Robustification, a promising approach for reducing processor power, converts applications into numerical optimization problems and solves them using gradient descent and conjugate gradient algorithms [1]. The improvement in robustness, however, comes at the expense of performance when compared to the baseline non-iterative versions of these applications. To mitigate the performance loss from robustification, we present the design of a hardware accelerator and corresponding software support that accelerate gradient descent and conjugate gradient based iterative implementation of applications. Unlike traditional accelerators, our design accelerates different types of linear algebra operations found in many algorithms and is capable of efficiently handling sparse matrices that arise in applications such as graph matching. We show that the proposed accelerator can provide significant speedups for iterative versions of several applications and that for some applications such as least squares, it can substantially improve the computation time as compared to the baseline non-iterative implementation.","PeriodicalId":375788,"journal":{"name":"2011 IEEE 9th Symposium on Application Specific Processors (SASP)","volume":"205 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2011-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"9","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2011 IEEE 9th Symposium on Application Specific Processors (SASP)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/SASP.2011.5941086","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 9

Abstract

Application Robustification, a promising approach for reducing processor power, converts applications into numerical optimization problems and solves them using gradient descent and conjugate gradient algorithms [1]. The improvement in robustness, however, comes at the expense of performance when compared to the baseline non-iterative versions of these applications. To mitigate the performance loss from robustification, we present the design of a hardware accelerator and corresponding software support that accelerate gradient descent and conjugate gradient based iterative implementation of applications. Unlike traditional accelerators, our design accelerates different types of linear algebra operations found in many algorithms and is capable of efficiently handling sparse matrices that arise in applications such as graph matching. We show that the proposed accelerator can provide significant speedups for iterative versions of several applications and that for some applications such as least squares, it can substantially improve the computation time as compared to the baseline non-iterative implementation.
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
梯度下降和共轭梯度的硬件加速技术
应用鲁棒化是一种很有前途的降低处理器功耗的方法,它将应用转化为数值优化问题,并使用梯度下降和共轭梯度算法求解[1]。然而,与这些应用程序的基线非迭代版本相比,健壮性的改进是以牺牲性能为代价的。为了减轻鲁棒化带来的性能损失,我们设计了一个硬件加速器和相应的软件支持,以加速梯度下降和基于共轭梯度的应用迭代实现。与传统的加速器不同,我们的设计加速了许多算法中发现的不同类型的线性代数运算,并且能够有效地处理图匹配等应用中出现的稀疏矩阵。我们表明,所提出的加速器可以为几个应用程序的迭代版本提供显著的加速,并且对于一些应用程序(如最小二乘),与基线非迭代实现相比,它可以大大提高计算时间。
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
相关文献
Russia: A Long View
IF 0.2 4区 社会学European Legacy-Toward New ParadigmsPub Date : 2015-04-13 DOI: 10.1080/10848770.2015.1028015
R. Sakwa
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
TARCAD: A template architecture for reconfigurable accelerator designs USHA: Unified software and hardware architecture for video decoding Frameworks for GPU Accelerators: A comprehensive evaluation using 2D/3D image registration System integration of Elliptic Curve Cryptography on an OMAP platform A hardware acceleration technique for gradient descent and conjugate gradient
×
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