Design of adder and MAX unit for high accuracy stochastic computing

ZHOU You; LI Jie; HE Guanghui

Chinese

您当前的位置：

首页 >

文章列表页 >

Design of adder and MAX unit for high accuracy stochastic computing

更新时间：2026-09-04

- Design of adder and MAX unit for high accuracy stochastic computing
- Vol. 38, Issue 7, Pages: 1-6(2021)
- 作者机构：
  
  1. 上海交通大学微纳电子学系,上海,200240
  2. 上海航天测控通信研究所,上海,200092
- 作者简介：
- 基金信息：
- DOI：
  CLC：
- Published：2021
- 稿件说明：
移动端阅览
周游, 李杰, 贺光辉. 面向高精度随机计算单元的加法运算电路与MAX运算电路设计[J]. 微电子学与计算机, 2021, 38(7): 1-6.

ZHOU You, LI Jie, HE Guanghui. Design of adder and MAX unit for high accuracy stochastic computing[J]. 2021, 38(7): 1-6.
周游, 李杰, 贺光辉. 面向高精度随机计算单元的加法运算电路与MAX运算电路设计[J]. 微电子学与计算机, 2021, 38(7): 1-6. DOI：

ZHOU You, LI Jie, HE Guanghui. Design of adder and MAX unit for high accuracy stochastic computing[J]. 2021, 38(7): 1-6. DOI：

摘要

随机计算(Stochastic Computing

SC)是一种以概率编码数值的新型技术，可以有效减小电路功耗和面积，但其所固有的随机特性会导致计算精度损失，从而限制了其应用范围.为提高随机计算的准确性，提出了基于误差补偿原理的高精度随机计算加法运算单元，设计了一种计算随机序列最大值的SC-MAX单元.仿真结果显示：与同类设计相比，两种计算单元的计算精度明显提高，其中同等SC序列长度下加法单元的相对误差降低了80%以上，SC-MAX计算单元的相对误差减小了90%，且LUT与FF开销减小了10%以上.将两种计算单元运用于神经网络，在MNIST及CIFAR10数据集上的推理结果显示：精度损失跟浮点数相比分别可低至0.04%和0.86%.

Abstract

Stochastic Computing(SC) is a new technology that encodes numerical value by probability

which can effectively reduce the power consumption and area of the circuit. However

the loss of calculation accuracy caused by its inherent random characteristics limits its application. In order to improve the accuracy of Stochastic Computing

a high-precision Stochastic Computing addition unit based on error compensation principle is proposed

and a SC-MAX element is designed to calculate the maximum value of SC sequence. The simulation results show that

compared with the similar design

the calculation accuracy of the two calculation units is significantly improved. The relative error of the addition unit is reduced by more than 80% under the same SC length

and the relative error of the SC-MAX calculation unit is reduced by 90% compared with the similar design

and the LUT and FF overhead is significantly reduced by more than 10%. The two kinds of calculation units are applied to neural network. The reasoning results on MNIST and CIFAR10 data sets show that the precision loss is less than 0.04%and 0.86% respectively comparedto float point results.

关键词

Keywords

references

REN A, LI Z, DING C W, et al. SC-DCNN: highly-scalable deep convolutional neural network using stochastic computing[C]//Proceedings of the Twenty-Second International Conference on Architectural Support for Programming Languages and Operating Systems. Virtual, MI, USA: Association for Computing Machinery, 2017: 405-418. DOI:10.1145/3037697.3037746.

AKHTAR R, KHANDAY F A. Stochastic computing: systems, applications, challenges and solutions[C]//2018 3rd International Conference on Communication and Electronics Systems (ICCES). Coimbatore, India: IEEE, 2018: 722-727. DOI:10.1109/CESYS.2018.8723960.

ALAGHI A, HAYES J P. Survey of stochastic computing[J]. ACM Transactions on Embedded Computing Systems, 2013, 12(2S): 92. DOI:10.1145/2465787.2465794.

KIM K, LEE J, CHOI K. Approximate de-randomizer for stochastic circuits[C]//2015 International SoC Design Conference (ISOCC). Gyeongju, Korea (South): IEEE, 2015: 123-124. DOI:10.1109/ISOCC.2015.7401667.

XIONG H F, ABU BAKAR M, HE G H. Hardware implementation of an improved stochastic computing based deep neural network using short sequence length[J]. IEEE Transactions on Circuits and Systems Ⅱ: Express Briefs, 2020, 67(11): 2667-2671. DOI:10.1109/TCSII.2020.2969691.

YU J, KIM K, LEE J, et al. Accurate and efficient stochastic computing hardware for convolutional neural networks[C]//2017 IEEE International Conference on Computer Design (ICCD). Boston, MA: IEEE, 2017: 105-112. DOI:10.1109/ICCD.2017.24.

熊海翡, 贺光辉. 基于随机计算的大规模MIMO检测算法研究与硬件实现[J]. 微电子学与计算机, 2020, 37(7): 36-41. DOI:10.19304/j.chki.issn1000-7180.2020.07.008. XIONG H F, HE G H. Design and implementation of stochastic computing based massive MIMO detector[J]. Microelectronics % Computer, 2020, 37(7): 36-41. DOI:10.19304/j.chki.issn1000-7180.2020.07.008.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Aerial vehicle detection algorithm based on multi-scale fusion and small-object enhancement

High gain stability high temperature operational amplifier based on reverse body bias

Design of ultra-low power real-time clock circuits

Lightweight countermeasures against Spectre attacks on the branch prediction unit

Research on the design of blockchain-based AI-generated content traceability framework

Related Author

No data

Related Institution

No data

AI问答

Postal code：100079
Tel：（010）53879206 Email：tmw@bjxintong.com.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09082226号-64 京公网安备11010602201714号
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰