脉冲神经网络硬件互连系统的动态优先级仲裁策略

doi:10.3969/j.issn.0372-2112.2018.08.014

电子学报 ›› 2018, Vol. 46 ›› Issue (8): 1898-1905.DOI: 10.3969/j.issn.0372-2112.2018.08.014

脉冲神经网络硬件互连系统的动态优先级仲裁策略

刘俊秀¹, 黄星月¹, 罗玉玲¹, 曹弋²

1. 广西师范大学电子工程学院, 广西桂林 541004;
2. 萨里大学萨里商学院商业转型和企业可持续性发展系, 英国萨里 GU27XH

收稿日期:2017-07-27 修回日期:2017-11-02 出版日期:2018-08-25
- 通讯作者:
- 罗玉玲
- 作者简介:
- 刘俊秀 男,1984年生于黑龙江绥化.博士、广西师范大学电子工程学院讲师.研究方向为生物神经网络、类脑系统;黄星月 男,1992年生于河南开封.广西师范大学电子工程学院硕士研究生.研究方向为脉冲神经网络大规模互连架构.
- 基金资助:
- 国家自然科学基金 (No.61603104，No.61661008）; 广西自然科学基金 (No.2016GXNSFCA380017，No.2015GXNSFBA139256）; 广西高校引进海外高层次人才百人计划专项经费; 广西高校科研项目 (No.KY2016YB059）; 广西多源信息挖掘与安全重点实验室开发基金 (No.MIMS15-07）

Dynamic Priority Arbitration Strategy for Interconnections of Hardware Spiking Neural Networks

LIU Jun-xiu¹, HUANG Xing-yue¹, LUO Yu-ling¹, CAO Yi²

1. Faculty of Electronic Engineering, Guangxi Normal University, Guilin, Guangxi 541004, China;
2. Department of Business Transformation and Sustainable Enterprise, Surrey Business School, University of Surrey, Surrey GU27 XH, UK

Received:2017-07-27 Revised:2017-11-02 Online:2018-08-25 Published:2018-08-25
- Corresponding author:
- LUO Yu-ling
- Supported by:
- National Natural Science Foundation of China (No.61603104, No.61661008); Natural Science Foundation of Guangxi Zhuang Autonomous Region, China (No.2016GXNSFCA380017, No.2015GXNSFBA139256); Overseas High-level Talents Introduction Project of Colleges and Universities of Guangxi Zhuang Autonomous Region ("100-Talents Program") Program"); Scientific Research Project of Universities in Guangxi Zhuang Autonomous Reigion (No.KY2016YB059); Guangxi Key Laboratory Foundation for Multi-source Information Mining and Security (No.MIMS15-07)

摘要/Abstract

摘要： 本文基于EMBRACE脉冲神经网络硬件实现方案提出了一种片上网络路由器的动态优先级仲裁策略，来解决脉冲神经网络脉冲传输的交通负载非均衡问题.该方案使用二维网格片上网络系统实现神经元之间的互连通讯，其基于脉冲发送频率的动态优先级仲裁策略能够降低高频路径的平均延迟及系统丢包风险，提高系统工作稳定性.使用Noxim片上网络模拟器搭建实验平台，测试结果表明采用提出的动态优先级仲裁策略较轮询及固定优先级仲裁器，高频路径延迟平均降低32.33%和34.69%，降低幅度最大达到84.86%和86.20%.在90nm CMOS工艺下，提出的路由器架构硬件面积为213，471μm²，具有较好的扩展性.

关键词: 脉冲神经网络, 片上网络, 路由器, 仲裁器, 动态优先级

Abstract: Based on the EMBRACE,i.e.a hardware architecture for the SNN,a dynamic priority arbitration strategy for a Networks-on-Chip (NoC) router is proposed to solve the non-balance traffic load problem of the SNN system.A2DmeshNoC system is used to provide the interconnected communication for the neurons.The dynamic priority arbitration strategy is based on the spiking transmission frequencies and it can reduce the risk of the packet loss and the average delay of the paths with heavy traffics,and improve the reliability of system.The Noxim simulator is used to build the simulation platform.The experimental results show that compared to the round-robin and fixed priority arbitration schemes,the proposed dynamic priority arbitration strategy has an average of 32.33% and 34.69%(maximum 84.86% and 86.20%) decrease for the delays of paths with heavy traffics.Based on the 90nm CMOS technology,the hardware area is 213,471 μ m²,which demonstrates its scalability.

Key words: spiking neural networks, networks-on-chip, router, arbiter, dynamic priority

中图分类号:

TP183

刘俊秀, 黄星月, 罗玉玲, 等. 脉冲神经网络硬件互连系统的动态优先级仲裁策略[J]. 电子学报, 2018, 46(8): 1898-1905.

LIU Jun-xiu, HUANG Xing-yue, LUO Yu-ling, et al. Dynamic Priority Arbitration Strategy for Interconnections of Hardware Spiking Neural Networks[J]. Acta Electronica Sinica, 2018, 46(8): 1898-1905.

参考文献

[1] 蔺想红,王向文,张宁,等.脉冲神经网络的监督学习算法研究综述[J].电子学报,2015,43(3):577-586. LIN Xiang-hong,WANG Xiang-wen,ZHANG Ning,et al.Supervised learning algorithms for spiking neural networks:A review[J].Acta Electronica Sinica,2015,43(3):577-586.(in Chinese)
[2] 马立伟,孙义和.片上网络拓朴优化:在离散平面上布局与布线[J].电子学报,2007,35(5):906-911. MA Li-wei,SUN Yi-he.Network-on-chip topology optimizations:Floor-plan and routing on discrete plane[J].Acta Electronica Sinica,2007,35(5):906-911.(in Chinese)
[3] 杨盛光,李丽,高明伦,等.面向能耗和延时的NoC映射方法[J].电子学报,2008,36(5):937-942. YANG Sheng-guang,LI Li,GAO Ming-lun,et al.An energy and delay aware mapping method of NoC[J].Acta Electronica Sinica,2008,36(5):937-942.(in Chinese)
[4] BENINIL,MICHELID,GIOVANNI.Networks on chips:a new SoC paradigm[J].Computer,2002,35(1):70-78.
[5] FIRUZAN A,MODARRESSI M,DANESHTALAB M.Reconfigurable communication fabric for efficient implementation of neural networks[A].10th International Symposium on Reconfigurable and Communication-centric Systems-on-Chip[C].IEEE,2015.
[6] CARRILLO S,HARKIN J,MCDAID L J,et al.Scalable hierarchical network-on-chip architecture for spiking neural network hardware implementations[J].IEEE Transactions on Parallel and Distributed Systems,2013,24(12):2451-2461.
[7] LIU J,HARKIN J,MCDAID L,et al.Hierarchical Networks-on-Chip interconnect for astrocyte-neuron network hardwar[A].Artificial Neural Networks and Machine Learning[C].Springer International Publishing,2016.382-389.
[8] LIU J,HARKIN J,LIAM P,et al.Scalable Networks-on-chip interconnected architecture for astrocyte-neuron networks[J].IEEE Transactions on Circuits and Systems I:Regular Papers,2016,63(12):2290-2303.
[9] HARKIN J,MORGAN F,MCDAID L,et al.A reconfigurable and biologically inspired paradigm for computation using network-on-chip and spiking neural networks[J].International Journal of Reconfigurable Computing,2009,2009:1-13.
[10] THEOCHARIDES T,LINK G,VIJAYKRISHNAN N,et al.A generic reconfigurable neural network architecture implemented as a network on chip[A].IEEE International SOC Conference[C].IEEE,2004.191-194.
[11] SCHEMMEL J,FIERES J,MEIER K.Wafer-scale integration of analog neural networks[A].Proceedings of the International Joint Conference on Neural Networks[C].IEEE,2008.431-438.
[12] FURBER S B,TEMPLE S,BROWNA.High-performance computing for systems of spiking neurons[J].Engineering,2006,2:29-36.
[13] JIN X,LUJAN M,PLANA L A,et al.Modeling spiking neural networks on SpiNNaker[J].Computing in Science and Engineering,2010,12(5):91-97.
[14] FURBER S B,LESTER D R,PLANAL A,et al.Overview of the SpiNNaker system architecture[J].IEEE Transactions on Computers,2013,62(12):2454-2467.
[15] SEO J-S,BREZZO B,LIU Y,et al.A 45nm CMOS neurosdsdsdsdsmorphic chip with a scalable architecture for learning in networks of spiking neurons[A].IEEE Custom integrated Circuits Conference (CICC)[C].IEEE,2011.1-4.
[16] MEROLLA P A,ARTHUR J V,ALVAREZ-ICAZA R,et al.A million spiking-neuron integrated circuit with a scalable communication network and interface[J].Science,2014,345(6197):668-673.
[17] CARRILLO S,HARKIN J,MCDAID L,et al.Advancing interconnect density for spiking neural network hardware implementations using traffic-aware adaptive network-on-chip routers[J].Neural Networks,Elsevier Ltd,2012,33:42-57.
[18] IZHIKEVICH E M.Simple model of spiking neurons[J].IEEE Transactions on Neural Networks,2003,14(6):1569-1572.
[19] A.L.HODGKIN,A.F.HUXLEY.A quantitative description of membrane and its application to conduction and excitation in nerve[J].Journal of Physiology,1952,117(4):500-544.
[20] 蔺想红,张田文.指数突触电导IF神经元模型及事件驱动模拟策略[J].电子学报,2008,36(8):1495-1501. LIN Xiang-hong,ZHANG Tiang-wen.An integrate and fire neuron model with exponential synaptic conductances for event-driven simulation strategy[J].Acta Electronica Sinica,2008,36(8):1495-1501.(in Chinese)

脉冲神经网络硬件互连系统的动态优先级仲裁策略

Dynamic Priority Arbitration Strategy for Interconnections of Hardware Spiking Neural Networks

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	王佳辉, 魏雯婷, 王领, 顾华玺. 光片上网络中基于NSGA-II的多目标映射方法研究[J]. 电子学报, 2022, 50(10): 2372-2380.
[2]	欧阳一鸣, 张鹏, 王奇, 安鑫, 梁华国. WiNoC中交叉开关仲裁及路由算法设计[J]. 电子学报, 2021, 49(3): 518-526.
[3]	陈国军, 蔺想红, 王国恩, 王向文. 一种脉冲神经元监督学习的直接计算方法[J]. 电子学报, 2021, 49(2): 331-337.
[4]	欧阳一鸣, 贾博远, 李建华, 黄正峰, 梁华国. WiNoC中无线通信拥塞与故障感知的容错路由算法[J]. 电子学报, 2020, 48(4): 662-669.
[5]	欧阳一鸣, 李胜, 黄正峰, 梁华国, 安鑫. WiNoC中无线节点和无线链路级拥塞避免的高效路由器设计[J]. 电子学报, 2018, 46(4): 775-782.
[6]	欧阳一鸣, 钱昌, 杜高明, 黄正峰, 李建华, 梁华国. WiNoC中EF-ACK容错无线接口设计[J]. 电子学报, 2018, 46(11): 2561-2570.
[7]	张浩为, 谢军伟, 师俊朋, 宗彬锋, 盛川. 动态优先级下防空相控阵雷达在线交错调度算法[J]. 电子学报, 2018, 46(1): 55-60.
[8]	王伟, 周梦玲, 方芳, 郭二辉, 陈田, 刘军, 任福继. 片上网络路由器的测试及其外壳旁路故障的诊断[J]. 电子学报, 2017, 45(3): 638-643.
[9]	黎建华, 吴宁, 胡永良, 张肖强. 一种自适应的混合型无线NoC拓扑结构[J]. 电子学报, 2016, 44(6): 1420-1428.
[10]	欧阳一鸣, 何鑫城, 梁华国, 易茂祥, 杜高明, 安鑫. 针对路径故障与局部拥塞的NoC容错路由算法[J]. 电子学报, 2016, 44(4): 920-925.
[11]	欧阳一鸣, 孙成龙, 陈奇, 梁华国, 易茂祥, 黄正峰, 闫爱斌. 3D NoC关键通信部件容错方法研究综述[J]. 电子学报, 2016, 44(12): 3053-3063.
[12]	王一拙, 左琦, 计卫星, 王小军, 石峰. 访存与用户行为敏感的MPSoC应用映射[J]. 电子学报, 2015, 43(4): 631-638.
[13]	蔺想红, 王向文, 张宁, 马慧芳. 脉冲神经网络的监督学习算法研究综述[J]. 电子学报, 2015, 43(3): 577-586.
[14]	欧阳一鸣, 陈义军, 梁华国, 易茂祥, 李建华. 一种故障通道隔离的低开销容错路由器设计[J]. 电子学报, 2014, 42(11): 2142-2149.
[15]	高建良, 李欣, 王建新. 并发追踪数据流的多缓存选址算法[J]. 电子学报, 2014, 42(11): 2310-2313.