Architecture design of re-configurable convolutional neural network on software definition

Peijie LI; Li ZHANG; Yunfei XIA; Liming XU

doi:10.11959/j.issn.2096-109x.2021043

您当前的位置：

首页 >

文章列表页 >

Architecture design of re-configurable convolutional neural network on software definition

TopicⅠ: Application of neural network technology | 更新时间：2024-06-03

- Architecture design of re-configurable convolutional neural network on software definition
- Chinese Journal of Network and Information Security Vol. 7, Issue 3, Pages: 29-36(2021)
- 作者机构：
  
  1. 信息工程大学，河南郑州 450001
  2. 天津市滨海新区信息技术创新中心，天津 300457
- 作者简介：
- 基金信息：
- DOI：10.11959/j.issn.2096-109x.2021043
  CLC： TP393
- Online First：2021-06，
  
  Published：15 June 2021
- 稿件说明：
移动端阅览
Peijie LI, Li ZHANG, Yunfei XIA, et al. Architecture design of re-configurable convolutional neural network on software definition[J]. Chinese Journal of Network and Information Security, 2021, 7(3): 29-36.
DOI：

Peijie LI, Li ZHANG, Yunfei XIA, et al. Architecture design of re-configurable convolutional neural network on software definition[J]. Chinese Journal of Network and Information Security, 2021, 7(3): 29-36. DOI： 10.11959/j.issn.2096-109x.2021043.

摘要

为满足卷积神经网络业务处理的灵活性和高性能需求，提出一种基于软件定义的可重构卷积神经网络架构。该架构采用归一化处理流程实现卷积层网络的动态重构与运算模式的加速。采用AHB和AXI的双总线架构，实现卷积神经网络的流水计算。通过软件定义在FPGA上实现了不同网络结构下的数据集实时处理。实验结果表明，所设计的FPGA电路能够实现两种网络模型的软件定义，网络模型与输入数据集相同的条件下，该架构的运算处理能力为CPU的10倍，运算能耗比为GPU的2倍。

Abstract

In order to meet the flexibility and efficiency requirement in convolutional neural network (CNN)

an architecture of re-configurable CNN based on software definition was proposed.In the architecture

the process of CNN could be normalized and the operation mode could be accelerated.The calculation pipeline was implemented by using dual bus architecture based on AHB and AXI protocols.By software definition

the proposed architecture

which could realize the real-time processing of data among different CNN structure

was implemented on FPGA.The result shows that at least 2 CNN models can be software defined on the FPGA circuit.The output measures an operation processing capacity of 10 times that of CPU

and an operation energy consumption ratio of 2 times that of GPU.

关键词

Keywords

references

SINGH K , TIWARI S C , GUPTA M . A closed-loop ASIC design approach based on logical effort theory and artificial neural networks [J ] . Integration , 2019 , 69 : 10 - 22 .

LIU Z , DOU Y , JIANG J , et al . An FPGA-based processor for training convolutional neural networks [C ] // International Conference on Field Programmable Technology (ICFPT) . 2018 .

方睿 , 刘加贺 , 薛志辉 , 等 . 卷积神经网络的 FPGA 并行加速方案设计 [J ] . 计算机工程与应用 , 2015 , 51 ( 8 ): 32 - 36 .

FANG R , LIU J H , XUE Z H , et al . FPGA-based design for convolution neural network [J ] . Computer Engineering and Applications , 2015 , 51 ( 8 ): 32 - 36 .

王巍 , 周凯利 , 王伊昌 , 等 . 卷积神经网络(CNN)算法的 FPGA并行结构设计 [J ] . 微电子学与计算机 , 2019 , 36 ( 4 ): 57 - 62 , 66 .

WANG W , ZHOU K L , WANG Y C , et al . FPGA parallel structure design of convolutional neural network(CNN) algorithm [J ] . Microelectronics ＆ Computer , 2019 , 36 ( 4 ): 57 - 62 .

翟社平 , 邱程 , 杨媛媛 , 等 . 基于 FPGA 的卷积神经网络加速器设计与实现 [J ] . 微电子学与计算机 , 2019 , 36 ( 8 ): 83 - 86 .

ZHAI S P , QIU C , YANG Y Y , et al . Design and implementation of convolutional neural network accelerator based on FPGA [J ] . Microelectronics ＆ Computer , 2019 , 36 ( 8 ): 83 - 86 .

窦阳 , 卿粼波 , 何小海 , 等 . 基于FPGA的CNN加速器设计与实现 [J ] . 信息技术与网络安全 , 2019 , 38 ( 11 ): 96 - 101 .

DOU Y , QING L B , HE X H , et al . Design and implementation of CNN accelerator based on FPGA [J ] . Information Technology and Network Security , 2019 , 38 ( 11 ): 96 - 101 .

周瑛 , 张铃 . 模糊集方法在检索评价系统中的应用 [J ] . 计算机技术与发展 , 2007 , 17 ( 1 ): 111 - 113 .

ZHOU Y , ZHANG L . Application of fuzzy measure in information retrieval evaluation [J ] . Computer Technology and Development , 2007 , 17 ( 1 ): 111 - 113 .

VENUGOPAL S , CASTRO-PAREJA C R , DANDEKAR O S . An FPGA-based 3D image processor with median and convolution filters for real-time applications [C ] // Proc of Spie-is＆t Electronic Imaging . 2005 .

CHEN J , PRODIC A , ERICKSON R W , et al . Predictive digital current programmed control [J ] . IEEE Transactions on Power Electronics , 2003 , 18 ( 1 ): 411 - 419 .

CHELLAPILLA K , PURI S , SIMARD P . High performance convolutional neural networks for document processing [C ] // 10th International Workshop on Frontiers in Handwriting Recognition . 2006 .

YU Y , WU C , ZHAO T , et al . OPU:an FPGA-based overlay processor for convolutional neural networks [J ] . IEEE Transactions on Very Large-Scale Integration (VLSI) Systems , 2019 , 28 ( 1 ): 35 - 47 .

WU D , PIGOU L , KINDERMANS P J , et al . Deep dynamic neural networks for multimodal gesture segmentation and recognition [J ] . IEEE Transactions on Pattern Analysis ＆ Machine Intelligence , 2016 , 38 ( 8 ): 1583 - 1597 .

仇越 , 马文涛 , 柴志雷 . 一种基于FPGA的卷积神经网络加速器设计与实现 [J ] . 微电子学与计算机 , 2018 , 35 ( 8 ): 68 - 72 .

QIU Y , MA W T , CHAI Z L . Design and implementation of a convolutional neural network accelerator based on FPGA [J ] . Microelectronics ＆ Computer , 2018 , 34 ( 8 ): 68 - 72 .

刘勤让 , 刘崇阳 , 周俊 , 等 . 基于线性脉动阵列的卷积神经网络计算优化与性能分析 [J ] . 网络与信息安全学报 , 2018 , 4 ( 2 ): 16 - 24 .

LIU Q R , LIU C Y , ZHOU J , et al . Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysis [J ] . Chinese Journal of Network and Information Security , 2018 , 4 ( 2 ): 16 - 24 .

梁爽 . 可重构神经网络加速器设计关键技术研究 [D ] . 北京：清华大学 , 2017 .

LIANG S . Research on key technologies of reconfigurable neural network accelerator design [D ] . Beijing:Tsinghua University , 2017 .

Views

238

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Implementation of SM3 algorithm based on SoPC component

Unbalanced protocol recognition method based on improved residual U-Net

High speed national secret SM4 optical fiber communication system scheme based on FPGA

Research on the robustness of convolutional neural networks in image recognition

Mining behavior pattern of mobile malware with convolutional neural network

Related Author

Xin-chun YIN

Yan-mei LI

Meng-li SHAO

Tiantian MA

Zheng HONG

Jisheng WU

Yangfan JIA

Jiabo SONG

Related Institution

School of Information Engineering, Yangzhou University

School of Guangling,Yangzhou University

⁰