MF-Conv: A Novel Convolutional Approach Using Bit-Resolution-based Weight Decomposition to Eliminate Multiplications for CNN Acceleration

Chen Yang, Xianxian Lv, Bowen Li, Shiquan Fan, K. Mei, Li Geng
{"title":"MF-Conv: A Novel Convolutional Approach Using Bit-Resolution-based Weight Decomposition to Eliminate Multiplications for CNN Acceleration","authors":"Chen Yang, Xianxian Lv, Bowen Li, Shiquan Fan, K. Mei, Li Geng","doi":"10.1109/ICSICT49897.2020.9278019","DOIUrl":null,"url":null,"abstract":"Convolution computation is the core of convolutional neural network (CNN). With the increasing demand for the accuracy of CNN applications, the amount of convolution computation has been increasing rapidly. Now, most FPGA-based CNN accelerators tend to utilize multiply-and-accumulate (MAC) arrays in convolution operations, whose DSP amount determines the computational roof. To elevate the roof, this paper proposed a Multiplication-Free Convolution (MF-Conv) scheme for convolution layers. MF-Conv utilizes a bit-resolution-based weight decomposition method to transform multiplications into additions. Hence, we can completely eliminate multiple operation in convolution computation, as a result, avoiding the usage of DSP. Experimental results showed that the implementation of MF-Conv on Xilinx XC7Z100 platform can run at a clock frequency of 279MHz. Moreover, Compared to ABM-SpConv, proposed MF-Conv improve the performance of 3x3 kernel by 9x. MF-Conv also has a much smaller hardware overhead compared with ABM-SpConv.","PeriodicalId":6727,"journal":{"name":"2020 IEEE 15th International Conference on Solid-State & Integrated Circuit Technology (ICSICT)","volume":"5 1","pages":"1-3"},"PeriodicalIF":0.0000,"publicationDate":"2020-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2020 IEEE 15th International Conference on Solid-State & Integrated Circuit Technology (ICSICT)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICSICT49897.2020.9278019","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1

Abstract

Convolution computation is the core of convolutional neural network (CNN). With the increasing demand for the accuracy of CNN applications, the amount of convolution computation has been increasing rapidly. Now, most FPGA-based CNN accelerators tend to utilize multiply-and-accumulate (MAC) arrays in convolution operations, whose DSP amount determines the computational roof. To elevate the roof, this paper proposed a Multiplication-Free Convolution (MF-Conv) scheme for convolution layers. MF-Conv utilizes a bit-resolution-based weight decomposition method to transform multiplications into additions. Hence, we can completely eliminate multiple operation in convolution computation, as a result, avoiding the usage of DSP. Experimental results showed that the implementation of MF-Conv on Xilinx XC7Z100 platform can run at a clock frequency of 279MHz. Moreover, Compared to ABM-SpConv, proposed MF-Conv improve the performance of 3x3 kernel by 9x. MF-Conv also has a much smaller hardware overhead compared with ABM-SpConv.
MF-Conv:一种新颖的卷积方法,使用基于位分辨率的权重分解来消除CNN加速的乘法
卷积计算是卷积神经网络(CNN)的核心。随着对CNN应用精度要求的不断提高,卷积计算量也在迅速增加。现在,大多数基于fpga的CNN加速器倾向于在卷积运算中使用乘法累加(MAC)数组,其DSP数量决定了计算量。为了解决这个问题,本文提出了一种卷积层的无乘法卷积(MF-Conv)方案。MF-Conv利用基于位分辨率的权重分解方法将乘法变换为加法。因此,我们可以完全消除卷积计算中的多次运算,从而避免使用DSP。实验结果表明,在Xilinx XC7Z100平台上实现的MF-Conv可以在279MHz的时钟频率下运行。此外,与ABM-SpConv相比,本文提出的MF-Conv将3x3内核的性能提高了9倍。与ABM-SpConv相比,MF-Conv的硬件开销也小得多。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信