Unicorn: A Multicore Neuromorphic Processor with Flexible Fan-In and Unconstrained Fan-Out for Neurons

被引：5

作者：

Yang, Zhijie ^{[1
]}

Wang, Lei ^{[1
]}

Wang, Yao ^{[1
]}

Peng, Linghui ^{[1
]}

Chen, Xiaofan ^{[1
]}

Xiao, Xun ^{[1
]}

Wang, Yaohua ^{[1
]}

Xu, Weixia ^{[1
]}

机构：

[1] Natl Univ Def Technol, Coll Comp Sci & Technol, Changsha, Hunan, Peoples R China

来源：

PROCEEDINGS OF THE 59TH ACM/IEEE DESIGN AUTOMATION CONFERENCE, DAC 2022 | 2022年

基金：

国家重点研发计划;

关键词：

multicore architecture; neuromorphic processor; spiking neural network; dynamic vision sensor; hardware accelerator; LARGE-SCALE MODEL;

D O I：

10.1145/3489517.3530563

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Neuromorphic processor is popular due to its high energy efficiency for spatio-temporal applications. However, when running the spiking neural network (SNN) topologies with the ever-growing scale, existing neuromorphic architectures face challenges due to their restrictions on neuron fan-in and fan-out. This paper proposes Unicorn, a multicore neuromorphic processor with a spike train sliding multicasting mechanism (STSM) and neuron merging mechanism (NMM) to support unconstrained fan-out and flexible fan-in of neurons. Unicorn supports 36K neurons and 45M synapses and thus supports a variety of neuromorphic applications. The peak performance and energy efficiency of Unicorn reach 36TSOPS and 424GSOPS/W respectively. Experimental results show that Unicorn can achieve 2x-5.5x energy reduction over the state-of-the-art neuromorphic processor when running an SNN with a relatively large fan-out and fan-in.

引用

页码：943 / 948

页数：6

共 50 条

[1] Fan-out and fan-in properties of superconducting neuromorphic circuits
Schneider, M. L.
Segall, K.
JOURNAL OF APPLIED PHYSICS, 2020, 128 (21)
[2] FAN-IN AND FAN-OUT WITH OPTICAL INTERCONNECTIONS
GOODMAN, JW
OPTICA ACTA, 1985, 32 (12): : 1489 - 1496
[3] Broadband Low-loss Fan-in/Fan-out Devices for Multicore Fibers
Kang, Yifan
Guo, Xiancong
Gan, Lin
Shen, Li
Yang, Chen
Zhang, Rui
Shen, Lei
Tong, Weijun
Fu, Songnian
Tang, Ming
Liu, Deming
2019 ASIA COMMUNICATIONS AND PHOTONICS CONFERENCE (ACP), 2019,
[4] Multicore Fibre Fan-In/Fan-Out Device using Fibre Optic Collimators
Jung, Y.
Hayes, J. R.
Alam, S. U.
Richardson, D. J.
43RD EUROPEAN CONFERENCE ON OPTICAL COMMUNICATION (ECOC 2017), 2017,
[5] Fan-in/fan-out device employing v-groove substrate for multicore fibre
Abe, Y.
Shikama, K.
Ono, H.
Yanagi, S.
Takahashi, T.
ELECTRONICS LETTERS, 2015, 51 (17) : 1347 - +
[6] Single-mode polymer waveguide for fan-in/fan-out of multicore fiber by the mosquito method
Yakabe, Sho
Ishigure, Takaaki
OPTICAL INTERCONNECTS XXII, 2022, 12007
[7] Multicore-Fiber Receptacle With Compact Fan-In/Fan-Out Device for SDM Transceiver Applications
Shikama, Kota
Abe, Yoshiteru
Kishi, Toshiki
Takeda, Koji
Fujii, Takuro
Nishi, Hidetaka
Matsui, Takashi
Aratake, Atsushi
Nakajima, Kazuhide
Matsuo, Shinji
JOURNAL OF LIGHTWAVE TECHNOLOGY, 2018, 36 (24) : 5815 - 5822
[8] An empirical study of fan-in and fan-out in Java OSS
Nasseri, E.
Counsell, S.
Tempero, E.
8th ACIS International Conference on Software Engineering Research, Management and Applications, SERA 2010, 2010, : 36 - 41
[9] Multicore-Fiber LC Receptacle with Compact Fan-in/Fan-out for Short-Reach Transceivers
Shikama, Kota
Abe, Yoshiteru
Kishi, Toshiki
Takeda, Koji
Fujii, Takuro
Nishi, Hidetaka
Matsui, Takashi
Aratake, Atsushi
Nakajima, Kazuhide
Matsuo, Shinji
2018 OPTICAL FIBER COMMUNICATIONS CONFERENCE AND EXPOSITION (OFC), 2018,
[10] Refactoring and its Relationship with Fan-in and Fan-out: An Empirical Study
Urgia, A. M.
Tonelli, R.
Marchesi, M.
Concas, G.
Counsell, S.
McFall, J.
Swift, S.
2012 16TH EUROPEAN CONFERENCE ON SOFTWARE MAINTENANCE AND REENGINEERING (CSMR), 2012, : 63 - 72

← 1 2 3 4 5 →