SqueezeFlow: A Sparse CNN Accelerator Exploiting Concise Convolution Rules

doi:10.1109/TC.2019.2924215

	SqueezeFlow: A Sparse CNN Accelerator Exploiting Concise Convolution Rules
	Li, Jiajun 1,2; Jiang, Shuhao 1,2; Gong, Shijun 1,2; Wu, Jingya 1,2; Yan, Junchao 1,2; Yan, Guihai 1,2; Li, Xiaowei 1,2
	2019-11-01
发表期刊	IEEE TRANSACTIONS ON COMPUTERS
ISSN	0018-9340
卷号	68 期号:11 页码:1663-1677
摘要	Convolutional Neural Networks (CNNs) have been widely used in machine learning tasks. While delivering state-of-the-art accuracy, CNNs are known as both compute- and memory-intensive. This paper presents the SqueezeFlow accelerator architecture that exploits sparsity of CNN models for increased efficiency. Unlike prior accelerators that trade complexity for flexibility, SqueezeFlow exploits concise convolution rules to benefit from the reduction of computation and memory accesses as well as the acceleration of existing dense architectures without intrusive PE modifications. Specifically, SqueezeFlow employs a PT-OS-sparse dataflow that removes the ineffective computations while maintaining the regularity of CNN computations. We present a full design down to the layout at 65 nm, with an area of 4.80mm2 and power of 536.09mW. The experiments show that SqueezeFlow achieves a speedup of 2:9 on VGG16 compared to the dense architectures, with an area and power overhead of only 8.8 and 15.3 percent, respectively. On three representative sparse CNNs, SqueezeFlow improves the performance and energy efficiency by 1:8 and 1:5 over the state-of-the-art sparse accelerators.
关键词	Convolutional neural networks accelerator architecture hardware acceleration
DOI	10.1109/TC.2019.2924215
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[61872336] ; National Natural Science Foundation of China[61532017] ; National Natural Science Foundation of China[61572470] ; National Natural Science Foundation of China[61432017] ; National Natural Science Foundation of China[61521092] ; National Natural Science Foundation of China[61376043] ; Youth Innovation Promotion Association, CAS[Y404441000]
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Hardware & Architecture ; Engineering, Electrical & Electronic
WOS记录号	WOS:000491426600009
出版者	IEEE COMPUTER SOC
引用统计	被引频次：32[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/2XEOYT63/14911
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Yan, Guihai; Li, Xiaowei
作者单位	1.Chinese Acad Sci, Inst Comp Technol, State Key Lab Comp Architecture, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Li, Jiajun,Jiang, Shuhao,Gong, Shijun,et al. SqueezeFlow: A Sparse CNN Accelerator Exploiting Concise Convolution Rules[J]. IEEE TRANSACTIONS ON COMPUTERS,2019,68(11):1663-1677.
APA	Li, Jiajun.,Jiang, Shuhao.,Gong, Shijun.,Wu, Jingya.,Yan, Junchao.,...&Li, Xiaowei.(2019).SqueezeFlow: A Sparse CNN Accelerator Exploiting Concise Convolution Rules.IEEE TRANSACTIONS ON COMPUTERS,68(11),1663-1677.
MLA	Li, Jiajun,et al."SqueezeFlow: A Sparse CNN Accelerator Exploiting Concise Convolution Rules".IEEE TRANSACTIONS ON COMPUTERS 68.11(2019):1663-1677.