基于FPGA集群的脉冲神经网络仿真器设计

doi:10.19678/j.issn.1000-3428.0056430

计算机工程 ›› 2020, Vol. 46 ›› Issue (10): 201-209. doi: 10.19678/j.issn.1000-3428.0056430

基于FPGA集群的脉冲神经网络仿真器设计

李康¹, 张鲁飞², 张新伟¹, 郁龚健¹, 刘家航¹, 吴东², 柴志雷^1,2

1. 江南大学物联网工程学院, 江苏无锡 214122;
2. 数学工程与先进计算国家重点实验室, 江苏无锡 214215

收稿日期:2019-10-28 修回日期:2019-12-09 发布日期:2019-12-23
作者简介:李康(1994-),男,硕士研究生,主研方向为计算机体系结构、FPGA技术;张鲁飞,工程师、博士;张新伟、郁龚健、刘家航,硕士研究生;吴东,研究员、博士;柴志雷,副教授、博士。
基金资助:
国家自然科学基金（61972180）；数学工程与先进计算国家重点实验室开放基金（2018A04）。

Design of Spiking Neural Network Simulator Based on FPGA Cluster

LI Kang¹, ZHANG Lufei², ZHANG Xinwei¹, YU Gongjian¹, LIU Jiahang¹, WU Dong², CHAI Zhilei^1,2

1. School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China;
2. State Key Laboratory of Mathematical Engineering and Advanced Computing, Wuxi, Jiangsu 214215, China

Received:2019-10-28 Revised:2019-12-09 Published:2019-12-23

摘要/Abstract

摘要： 针对类脑计算系统中NEST脉冲神经网络仿真器运行速度慢和功耗高的问题，设计一种基于现场可编程逻辑门阵列（FPGA）集群的NEST脉冲神经网络仿真器。在改进NEST仿真器结构的基础上，提出漏电流整合放电神经元计算模块的流水线并行架构，实现支持双核双线程和多节点多进程的FPGA集群设计。在皮质层视觉仿真模型上的实验结果表明，与基于Xeon E5-2620和ARM A9平台的NEST仿真器相比，基于FPGA集群的NEST仿真器计算能效和速度分别提升43.93倍、23.54倍和12.36倍、208倍，能为大规模类脑计算系统实现提供技术支持。

关键词: 类脑计算系统, 脉冲神经网络仿真器, 现场可编程逻辑门阵列集群, 硬件实现, 皮质层视觉仿真模型

Abstract: To address the low running speed and high power consumption of the NEST Spiking Neural Network(SNN) simulators in the brain-like computing systems,this paper proposes a NEST simulator based on Field Programmable Gate Array(FPGA) cluster for SNN.By improving the structure of NEST simulator,a pipeline parallel architecture of the Leaky Integrate and Fire(LIF) neuron calculation module is proposed,which realizes the design of the dual-core dual-thread and multi-node multi-process FPGA cluster.The experimental results of cortical visual simulation model show that the computational energy efficiency of the proposed FPGA-cluster-based NEST simulator is 43.93 times that of the Xeon E5-2620 and 23.54 times that of the ARM A9.The computational speed of the proposed simulator is 12.36 times that of the Xeon E5-2620 and 208 times that of the ARM A9.

Key words: brain-like computing system, Spiking Neural Network(SNN) simulator, Field Programmable Gate Array(FPGA) cluster, hardware implementation, cortical visual simulation model

中图分类号:

TP183

李康, 张鲁飞, 张新伟, 郁龚健, 刘家航, 吴东, 柴志雷. 基于FPGA集群的脉冲神经网络仿真器设计[J]. 计算机工程, 2020, 46(10): 201-209.

LI Kang, ZHANG Lufei, ZHANG Xinwei, YU Gongjian, LIU Jiahang, WU Dong, CHAI Zhilei. Design of Spiking Neural Network Simulator Based on FPGA Cluster[J]. Computer Engineering, 2020, 46(10): 201-209.

https://www.ecice06.com/CN/Y2020/V46/I10/201

图/表 18

20201023093039

20201023093043

20201023093046

20201023093050

20201023093054

20201023093058

20201023093102

20201023093105

20201023093108

20201023093121

20201023093125

20201023093129

20201023093132

20201023093135

20201023093139

20201023093142

20201023093149

20201023093157

参考文献

[1] MARCUS G.Deep learning:a critical appraisal[EB/OL].[2019-09-04].https://arxiv.org/abs/1801.00631.
[2] FURBER S B,LESTER D R,PLANA L A,et al.Overview of the SpiNNaker system architecture[J].IEEE Transactions on Computers,2013,62(12):2454-2467.
[3] AMUNTS K,EBELL C,MULLER J,et al.The human brain project:creating a European research infrastructure to decode the human brain[J].Neuron,2016,92(3):574-581.
[4] POO M M.Where to the mega brain projects?[J].National Science Review,2014,1(1):12-14.
[5] PU Muming,XU Bo,TAN Tieniu.Brain science and brain-inspired intelligence technology-an overview[J].Bulletin of Chinese Academy of Sciences,2016,31(7):725-736.(in Chinese)蒲慕明,徐波,谭铁牛.脑科学与类脑研究概述[J].中国科学院院刊,2016,31(7):725-736.
[6] HUANG Tiejun,SHI Luping,TANG Huajin,et al.Research on multimedia technology 2015-advances and trend of brain-like computing[J].Journal of Image and Graphics,2018,21(11):1411-1424.(in Chinese)黄铁军,施路平,唐华锦,等.多媒体技术研究:2015——类脑计算的研究进展与发展趋势[J].中国图象图形学报,2018,21(11):1411-1424.
[7] ZHENG Nanning,REN Pengju,CHEN Badong,et al.Brain-like(brain-inspired) computing problems and visual cognition[C]//Proceedings of the 19th Annual Meeting of China Association for Science and Technology.Beijing:China Association for Science and Technology,2017:1-3.(in Chinese)郑南宁,任鹏举,陈霸东,等.类脑(受脑启发的)计算的问题与视觉认知[C]//第十九届中国科协年会论文集.北京:中国科学技术协会,2017:1-3.
[8] MAASS W.Networks of spiking neurons:the third generation of neural network models[J].Neural Networks,1997,10(9):1659-1671.
[9] GEWALTIG M O,DIESMANN M.NEST(neural simulation tool)[J].Scholarpedia,2007,2(4):1430-1438.
[10] STIMBERG M,GOODMAN D F,BENICHOUX V,et al.Brian 2-the second coming:spiking neural network simulation in Python with code generation[J].BMC Neuroscience,2013,14(S1):38-45.
[11] MIGLIORE M,CANNIA C,LYTTON W W,et al.Parallel network simulations with neuron[J].Journal of Computational Neuroscience,2006,21(2):119-129.
[12] KNIGHT J C,NOWOTNY T.GPUs outperform current HPC and neuromorphic solutions in terms of speed and energy when simulating a highly-connected cortical model[EB/OL].[2019-09-04].https://www.frontiersin.org/articles/10.3389/fnins.2018.00941/full.
[13] WU Qingxiang,LIAO Xiaodong,HUANG Xi,et al.Development of FPGA toolbox for implementation of spiking neural networks[C]//Proceedings of the 5th International Conference on Communication Systems and Network Technologies.Washington D.C.,USA:IEEE Press,2015:27-48.
[14] SCHUMAN C D,POTOK T E,PATTON R,et al.A survey of neuromorphic computing and neural networks in hardware[EB/OL].[2019-09-04].https://arxiv.org/abs/1705.06963.
[15] CHEUNG K,SCHULTZ S R,LUK W.NeuroFlow:a general purpose spiking neural network simulation platform using customizable processors[J].Frontiers in Neuroscience,2016,9:516-522.
[16] PODOBAS A,MATSUOKA S.Designing and accelerating spiking neural networks using OpenCL for FPGAs[C]//Proceedings of 2017 International Conference on Field Programmable Technology.Washington D.C.,USA:IEEE Press,2017:11-25.
[17] POTJANS T C,DIESMANN M.The cell-type specific cortical microcircuit:relating structure and activity in a full-scale spiking network model[J].Cerebral Cortex,2014,24(3):785-806.
[18] NGUYEN Q A P,ANDELFINGER P,CAI W T,et al.Transitioning spiking neural network simulators to heterogeneous hardware[C]//Proceedings of 2019 ACM SIGSIM Conference on Principles of Advanced Discrete Simulation.New York,USA:ACM Press,2019:11-23.
[19] MASQUELIER T,THORPE S J.Unsupervised learning of visual features through spike timing dependent plasticity[J].PLoS Computational Biology,2007,3(2):31-45.
[20] IZHIKEVICH E M.Which model to use for cortical spiking neurons?[J].IEEE Transactions on Neural Networks,2004,15(5):1063-1070.
[21] LEE D,LEE G,KWON D,et al.Flexon:a flexible digital neuron for efficient spiking neural network simulations[C]//Proceedings of the 45th Annual International Symposium on Computer Architecture.Washington D.C.,USA:IEEE Press,2018:275-288.

选择文件类型/文献管理软件名称

选择包含的内容

基于FPGA集群的脉冲神经网络仿真器设计

Design of Spiking Neural Network Simulator Based on FPGA Cluster

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献

相关文章 14

编辑推荐

Metrics

本文评价

[1]	朱文锋, 王琴, 郭筝, 刘军荣. 针对分组密码的攻击方法研究[J]. 计算机工程, 2020, 46(1): 102-107,113.
[2]	夏文涛, 潘森杉, 王良民. 一种面向RFID的超轻量级流密码算法[J]. 计算机工程, 2019, 45(10): 144-149.
[3]	郭泓键,董秀则,高献伟. 基于FPGA的祖冲之算法硬件实现[J]. 计算机工程, 2014, 40(8): 268-272.
[4]	董传盛, 曾真, 谈熙, 闵昊. 一种适合于硬件实现的数字预失真算法[J]. 计算机工程, 2012, 38(7): 230-232.
[5]	杨镇西, 张丽, 聂智良. 基于SVM概率建模的硬件实现优化算法[J]. 计算机工程, 2011, 37(23): 217-219.
[6]	杨同杰, 戴紫彬, 杨晓辉. 一种可重构模乘器的硬件设计[J]. 计算机工程, 2011, 37(13): 238-240.
[7]	蔡瑞;须文波;柴志雷;王斌;刘凡. 粒子群优化算法的硬件实现及其性能分析[J]. 计算机工程, 2010, 36(4): 168-168.
[8]	齐爱学, 王洪刚, 李卫兵. 开关超混沌系统及其电路实现[J]. 计算机工程, 2010, 36(13): 205-207.
[9]	薛小龙, 雷磊, 许宗泽. 硬件平台上的HEED实现及其性能分析[J]. 计算机工程, 2010, 36(13): 111-113.
[10]	刘洪波;蔡慧;韩国栋. 移动IP RO中的MOBEE设计与实现[J]. 计算机工程, 2009, 35(7): 90-92.
[11]	魏利华;丁辉;宣军英;刘小晶. 面向入侵检测的网络处理器设计[J]. 计算机工程, 2008, 34(20): 154-155.
[12]	何德彪;陈建华;孙金龙. RSA/ECC密码协处理器的硬件实现[J]. 计算机工程, 2007, 33(22): 29-31,3.
[13]	付晓丽;张晓彤;王沁;刘大力;刘文庆. 一种面积有效缩减的AES算法硬件实现[J]. 计算机工程, 2006, 32(22): 159-161.
[14]	何德彪;陈建华;胡进. GF(2m)上椭圆曲线密码协处理器的硬件实现[J]. 计算机工程, 2006, 32(19): 146-148.

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于FPGA集群的脉冲神经网络仿真器设计

Design of Spiking Neural Network Simulator Based on FPGA Cluster

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献

相关文章 14

编辑推荐

Metrics

本文评价