BSHIFT: A Low Cost Deep Neural Networks Accelerator

doi:10.1007/s10766-018-00624-9

	BSHIFT: A Low Cost Deep Neural Networks Accelerator
	Yu, Yong 1,2,3; Zhi, Tian 1; Zhou, Xuda 4; Liu, Shaoli 1,3; Chen, Yunji 1,2; Cheng, Shuyao 5
	2019-06-01
发表期刊	INTERNATIONAL JOURNAL OF PARALLEL PROGRAMMING
ISSN	0885-7458
卷号	47 期号:3 页码:360-372
摘要	Deep neural networks (DNNs) have become ubiquitous in artificial intelligence applications, including image processing, speech processing and natural language processing. However, the main characteristic of DNNs is that they are computationally and memory intensive, making them difficult to deploy on embedded systems with limited hardware resources and power budgets. To address this limitation, we introduce a new quantization method with mixed data structure and bit-shifting broadcast accelerator structure BSHIFT. These works together reduce the storage requirement of neural networks models from 32 to 5bits without affecting their accuracy. We implement BSHIFT at TSMC 16nm technology node, and the efficiency achieves 64TOPS/s per watt in our experiments.
关键词	Deep neural networks Low power Lossless Accelerator
DOI	10.1007/s10766-018-00624-9
收录类别	SCI
语种	英语
资助项目	National Key Research and Development Program of China[2017YFA0700902] ; National Key Research and Development Program of China[2017YFB1003101] ; NSF of China[61472396] ; NSF of China[61432016] ; NSF of China[61473275] ; NSF of China[61522211] ; NSF of China[61532016] ; NSF of China[61521092] ; NSF of China[61502446] ; NSF of China[61672491] ; NSF of China[61602441] ; NSF of China[61602446] ; NSF of China[61732002] ; NSF of China[61702478] ; 973 Program of China[2015CB358800] ; National Science and Technology Major Project[2018ZX01031102]
WOS研究方向	Computer Science
WOS类目	Computer Science, Theory & Methods
WOS记录号	WOS:000471644400003
出版者	SPRINGER/PLENUM PUBLISHERS
引用统计	被引频次：7[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/4194
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Yu, Yong
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 3.Cambricon Tech Ltd, Beijing 100191, Peoples R China 4.Univ Sci & Technol China, Hefei 230026, Anhui, Peoples R China 5.Tsinghua Univ, Beijing 100084, Peoples R China
推荐引用方式 GB/T 7714	Yu, Yong,Zhi, Tian,Zhou, Xuda,et al. BSHIFT: A Low Cost Deep Neural Networks Accelerator[J]. INTERNATIONAL JOURNAL OF PARALLEL PROGRAMMING,2019,47(3):360-372.
APA	Yu, Yong,Zhi, Tian,Zhou, Xuda,Liu, Shaoli,Chen, Yunji,&Cheng, Shuyao.(2019).BSHIFT: A Low Cost Deep Neural Networks Accelerator.INTERNATIONAL JOURNAL OF PARALLEL PROGRAMMING,47(3),360-372.
MLA	Yu, Yong,et al."BSHIFT: A Low Cost Deep Neural Networks Accelerator".INTERNATIONAL JOURNAL OF PARALLEL PROGRAMMING 47.3(2019):360-372.