中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 计算技术研究所
  3. 中国科学院计算技术研究所
  4. 中国科学院计算技术研究所期刊论文
  5. 英文
Swallow: A Versatile Accelerator for Sparse Neural Networks

文献类型：期刊论文

作者Liu, Bosheng1,2; Chen, Xiaoming1,2; Han, Yinhe1,2; Xu, Haobo1,2
刊名IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
出版日期2020-12-01
卷号39期号:12页码:4881-4893
ISSN号0278-0070
关键词Accelerator convolutional (Conv) layers fully connected (FC) layers sparse neural networks (SNNs)
DOI10.1109/TCAD.2020.2978836
英文摘要Sparse neural networks (SNNs) are emerging as a promising technique for resource-limited intelligent embedded systems because of the compact model size and the uncompromised accuracy. Recently, most of the dedicated neural network accelerators are beginning to exploit the sparsity of neural network models for performance boost and energy saving. However, existing sparsity-aware accelerators fail to support both sparse weights and activations in neural networks or support them at the same time for both convolutional (Conv) layers and fully connected (FC) layers, which dominate the computational time of neural networks. In this article, we propose a novel sparsity-aware accelerator architecture, called Swallow, to sufficiently improve the inference performance by eliminating ineffectual weights and activations of neural networks. Swallow comprises: 1) a 2-D systolic architecture that fully utilizes the sparsity of both weights and activations in both Conv and FC layers and 2) a sparsity-aware dataflow which is optimized to reuse both weights and activations and to achieve high processing element (PE) utilization by sparse matrix multiplication tiling. Comprehensive evaluations based on a place-and-route process show that Swallow, with 614 GOP/s peak performance and 1.26-W power, outperforms a state-of-the-art sparsity-aware accelerator Cambricon-X by 1.32x in term of energy efficiency.
资助项目National Key Research and Development Program of China[2018YFA0701500] ; Key Research Program of Frontier Sciences, CAS[ZDBS-LY-JSC012] ; National Natural Science Foundation of China[61804155] ; National Natural Science Foundation of China[61834006] ; Youth Innovation Promotion Association CAS ; Young Elite Scientists Sponsorship Program by CAST[2018QNRC001] ; Beijing Academy of Artificial Intelligence (BAAI)
WOS研究方向Computer Science ; Engineering
语种英语
出版者IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号WOS:000592111400045
源URL[http://119.78.100.204/handle/2XEOYT63/16138]  
专题中国科学院计算技术研究所期刊论文_英文
通讯作者Chen, Xiaoming
作者单位1.Chinese Acad Sci, Inst Comp Technol, Ctr Intelligent Comp Syst, 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		 Liu, Bosheng,Chen, Xiaoming,Han, Yinhe,et al. Swallow: A Versatile Accelerator for Sparse Neural Networks[J]. IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS,2020,39(12):4881-4893.
APA Liu, Bosheng,Chen, Xiaoming,Han, Yinhe,&Xu, Haobo.(2020).Swallow: A Versatile Accelerator for Sparse Neural Networks.IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS,39(12),4881-4893.
MLA Liu, Bosheng,et al."Swallow: A Versatile Accelerator for Sparse Neural Networks".IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS 39.12(2020):4881-4893.

入库方式： OAI收割

来源：计算技术研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。