An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks

Chen Yang; Haibo Zhang; Xiaoli Wang; Li Geng

doi:10.1109/ICSICT.2018.8565823

An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks

Chen Yang
, Haibo Zhang
, Xiaoli Wang
, Li Geng

Faculty of Electronic and Information Engineering

Xi'an Jiaotong University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

8 Scopus citations

Abstract

Multiple Convolutional Neural Networks (CNNs) become widely used in modern AI systems. There is increasingly necessity to apply different CNN shapes for different scenarios. However, it also brings challenges on throughput, energy efficiency and flexibility to hardware. In this paper, a novel accelerator, called reconfigurable neural accelerator (RNA), was proposed based on reconfigurable computing technology. In addition, image row broadcast (IRB) and zero detection technology (ZDT) were applied for increased energy efficiency and throughput. IRB can optimize the convolutional dataflow on spatial array architecture with 22×22 processing elements, increasing data reuse and reducing data movement. ZDT reduces the weight data access of the fully connected layer. At the cost of 10.25W power consumption on Virtex UltraScale XCVU440 platform, RNA can process the convolutional layers at 97.4 GOPS for AlexNet, at 90.75GOPS for VGG and at 100.8 GOPS for Lenet-5, respectively.

Original language	English
Title of host publication	2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings
Editors	Ting-Ao Tang, Fan Ye, Yu-Long Jiang
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781538644409
DOIs	https://doi.org/10.1109/ICSICT.2018.8565823
State	Published - 5 Dec 2018
Event	14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Qingdao, China Duration: 31 Oct 2018 → 3 Nov 2018

Publication series

Name	2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings

Conference

Conference	14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018
Country/Territory	China
City	Qingdao
Period	31/10/18 → 3/11/18

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy

Keywords

CNN
Image Row Broadcast dataflow
Reconfigurable computing
Zero Detection Technology

Access to Document

10.1109/ICSICT.2018.8565823

Cite this

Yang, C., Zhang, H., Wang, X., & Geng, L. (2018). An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks. In T.-A. Tang, F. Ye, & Y.-L. Jiang (Eds.), 2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings Article 8565823 (2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICSICT.2018.8565823

Yang, Chen ; Zhang, Haibo ; Wang, Xiaoli et al. / An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks. 2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings. editor / Ting-Ao Tang ; Fan Ye ; Yu-Long Jiang. Institute of Electrical and Electronics Engineers Inc., 2018. (2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings).

@inproceedings{618abdf0060542ceab286b792b38024f,

title = "An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks",

abstract = "Multiple Convolutional Neural Networks (CNNs) become widely used in modern AI systems. There is increasingly necessity to apply different CNN shapes for different scenarios. However, it also brings challenges on throughput, energy efficiency and flexibility to hardware. In this paper, a novel accelerator, called reconfigurable neural accelerator (RNA), was proposed based on reconfigurable computing technology. In addition, image row broadcast (IRB) and zero detection technology (ZDT) were applied for increased energy efficiency and throughput. IRB can optimize the convolutional dataflow on spatial array architecture with 22×22 processing elements, increasing data reuse and reducing data movement. ZDT reduces the weight data access of the fully connected layer. At the cost of 10.25W power consumption on Virtex UltraScale XCVU440 platform, RNA can process the convolutional layers at 97.4 GOPS for AlexNet, at 90.75GOPS for VGG and at 100.8 GOPS for Lenet-5, respectively.",

keywords = "CNN, Image Row Broadcast dataflow, Reconfigurable computing, Zero Detection Technology",

author = "Chen Yang and Haibo Zhang and Xiaoli Wang and Li Geng",

note = "Publisher Copyright: {\textcopyright} 2018 IEEE.; 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 ; Conference date: 31-10-2018 Through 03-11-2018",

year = "2018",

month = dec,

day = "5",

doi = "10.1109/ICSICT.2018.8565823",

language = "英语",

series = "2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

editor = "Ting-Ao Tang and Fan Ye and Yu-Long Jiang",

booktitle = "2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings",

}

Yang, C, Zhang, H, Wang, X & Geng, L 2018, An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks. in T-A Tang, F Ye & Y-L Jiang (eds), 2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings., 8565823, 2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018, Qingdao, China, 31/10/18. https://doi.org/10.1109/ICSICT.2018.8565823

An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks. / Yang, Chen; Zhang, Haibo; Wang, Xiaoli et al.
2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings. ed. / Ting-Ao Tang; Fan Ye; Yu-Long Jiang. Institute of Electrical and Electronics Engineers Inc., 2018. 8565823 (2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks

AU - Yang, Chen

AU - Zhang, Haibo

AU - Wang, Xiaoli

AU - Geng, Li

PY - 2018/12/5

Y1 - 2018/12/5

N2 - Multiple Convolutional Neural Networks (CNNs) become widely used in modern AI systems. There is increasingly necessity to apply different CNN shapes for different scenarios. However, it also brings challenges on throughput, energy efficiency and flexibility to hardware. In this paper, a novel accelerator, called reconfigurable neural accelerator (RNA), was proposed based on reconfigurable computing technology. In addition, image row broadcast (IRB) and zero detection technology (ZDT) were applied for increased energy efficiency and throughput. IRB can optimize the convolutional dataflow on spatial array architecture with 22×22 processing elements, increasing data reuse and reducing data movement. ZDT reduces the weight data access of the fully connected layer. At the cost of 10.25W power consumption on Virtex UltraScale XCVU440 platform, RNA can process the convolutional layers at 97.4 GOPS for AlexNet, at 90.75GOPS for VGG and at 100.8 GOPS for Lenet-5, respectively.

AB - Multiple Convolutional Neural Networks (CNNs) become widely used in modern AI systems. There is increasingly necessity to apply different CNN shapes for different scenarios. However, it also brings challenges on throughput, energy efficiency and flexibility to hardware. In this paper, a novel accelerator, called reconfigurable neural accelerator (RNA), was proposed based on reconfigurable computing technology. In addition, image row broadcast (IRB) and zero detection technology (ZDT) were applied for increased energy efficiency and throughput. IRB can optimize the convolutional dataflow on spatial array architecture with 22×22 processing elements, increasing data reuse and reducing data movement. ZDT reduces the weight data access of the fully connected layer. At the cost of 10.25W power consumption on Virtex UltraScale XCVU440 platform, RNA can process the convolutional layers at 97.4 GOPS for AlexNet, at 90.75GOPS for VGG and at 100.8 GOPS for Lenet-5, respectively.

KW - CNN

KW - Image Row Broadcast dataflow

KW - Reconfigurable computing

KW - Zero Detection Technology

UR - https://www.scopus.com/pages/publications/85060305393

U2 - 10.1109/ICSICT.2018.8565823

DO - 10.1109/ICSICT.2018.8565823

M3 - 会议稿件

AN - SCOPUS:85060305393

T3 - 2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings

BT - 2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings

A2 - Tang, Ting-Ao

A2 - Ye, Fan

A2 - Jiang, Yu-Long

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018

Y2 - 31 October 2018 through 3 November 2018

ER -

Yang C, Zhang H, Wang X, Geng L. An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks. In Tang TA, Ye F, Jiang YL, editors, 2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2018. 8565823. (2018 14th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2018 - Proceedings). doi: 10.1109/ICSICT.2018.8565823

An Energy-Efficient and Flexible Accelerator based on Reconfigurable Computing for Multiple Deep Convolutional Neural Networks

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this