Hardware Implementation of a Resource-Efficient Router for Multi-Core Spiking Neural Networks

Maryam Sadeghi; Yasser Rezaeiyan; Darío Fernández Khatiboun; Farshad Moradi

doi:10.1109/ISCAS46773.2023.10182040

Hardware Implementation of a Resource-Efficient Router for Multi-Core Spiking Neural Networks

Maryam Sadeghi, Yasser Rezaeiyan, Darío Fernández Khatiboun, Farshad Moradi

Department of Electrical and Computer Engineering - Electronics and Photonics

Research output: Contribution to book/anthology/report/proceeding › Article in proceedings › Research › peer-review

Abstract

Spiking neural networks (SNNs) are envisioned to be a better alternative to artificial neural networks (ANNs) for targeted applications. Multi-core implementation of SNNs has been built to achieve a resource-efficient design. However, managing the spike traffic congestion while routing the spikes between different cores requires a performance-resource tradeoff to avoid any packet loss. This paper presents a novel router architecture servicing ongoing packets in a 2-D mesh network while guaranteeing no packet drop. Here, the packets are distributed across different paths to reduce spike traffic. The proposed router suitable for a 16 × 16 network occupies an area of 0.001mm2 in 28nm CMOS technology, while consuming 75 fJ/transmission.

Original language	English
Title of host publication	ISCAS 2023 - 56th IEEE International Symposium on Circuits and Systems, Proceedings
Publisher	IEEE
Publication date	2023
Pages	1-5
ISBN (Print)	978-1-6654-5109-3
ISBN (Electronic)	9781665451093
DOIs	https://doi.org/10.1109/ISCAS46773.2023.10182040
Publication status	Published - 2023
Event	IEEE International Symposium on Circuits and Systems - Monterey, United States Duration: 21 May 2023 → 25 May 2023

Conference

Conference	IEEE International Symposium on Circuits and Systems
Country/Territory	United States
City	Monterey
Period	21/05/2023 → 25/05/2023

Keywords

Brain-inspired computing
artificial neural network
neuromorphic computing
spiking neural network

Access to Document

10.1109/ISCAS46773.2023.10182040

Cite this

@inproceedings{5ec50332dbb4408282818d6b737f298c,

title = "Hardware Implementation of a Resource-Efficient Router for Multi-Core Spiking Neural Networks",

abstract = "Spiking neural networks (SNNs) are envisioned to be a better alternative to artificial neural networks (ANNs) for targeted applications. Multi-core implementation of SNNs has been built to achieve a resource-efficient design. However, managing the spike traffic congestion while routing the spikes between different cores requires a performance-resource tradeoff to avoid any packet loss. This paper presents a novel router architecture servicing ongoing packets in a 2-D mesh network while guaranteeing no packet drop. Here, the packets are distributed across different paths to reduce spike traffic. The proposed router suitable for a 16 × 16 network occupies an area of 0.001mm2 in 28nm CMOS technology, while consuming 75 fJ/transmission.",

keywords = "Brain-inspired computing, artificial neural network, neuromorphic computing, spiking neural network",

author = "Maryam Sadeghi and Yasser Rezaeiyan and {Fern{\'a}ndez Khatiboun}, Dar{\'i}o and Farshad Moradi",

year = "2023",

doi = "10.1109/ISCAS46773.2023.10182040",

language = "English",

isbn = "978-1-6654-5109-3",

pages = "1--5",

booktitle = "ISCAS 2023 - 56th IEEE International Symposium on Circuits and Systems, Proceedings",

publisher = "IEEE",

note = "IEEE International Symposium on Circuits and Systems ; Conference date: 21-05-2023 Through 25-05-2023",

}

Sadeghi, M , Rezaeiyan, Y , Fernández Khatiboun, D & Moradi, F 2023, Hardware Implementation of a Resource-Efficient Router for Multi-Core Spiking Neural Networks. in ISCAS 2023 - 56th IEEE International Symposium on Circuits and Systems, Proceedings. IEEE, pp. 1-5, IEEE International Symposium on Circuits and Systems, Monterey, California, United States, 21/05/2023. https://doi.org/10.1109/ISCAS46773.2023.10182040

Hardware Implementation of a Resource-Efficient Router for Multi-Core Spiking Neural Networks. / Sadeghi, Maryam ; Rezaeiyan, Yasser ; Fernández Khatiboun, Darío et al.
ISCAS 2023 - 56th IEEE International Symposium on Circuits and Systems, Proceedings. IEEE, 2023. p. 1-5.

Research output: Contribution to book/anthology/report/proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Hardware Implementation of a Resource-Efficient Router for Multi-Core Spiking Neural Networks

AU - Sadeghi, Maryam

AU - Rezaeiyan, Yasser

AU - Fernández Khatiboun, Darío

AU - Moradi, Farshad

PY - 2023

Y1 - 2023

N2 - Spiking neural networks (SNNs) are envisioned to be a better alternative to artificial neural networks (ANNs) for targeted applications. Multi-core implementation of SNNs has been built to achieve a resource-efficient design. However, managing the spike traffic congestion while routing the spikes between different cores requires a performance-resource tradeoff to avoid any packet loss. This paper presents a novel router architecture servicing ongoing packets in a 2-D mesh network while guaranteeing no packet drop. Here, the packets are distributed across different paths to reduce spike traffic. The proposed router suitable for a 16 × 16 network occupies an area of 0.001mm2 in 28nm CMOS technology, while consuming 75 fJ/transmission.

AB - Spiking neural networks (SNNs) are envisioned to be a better alternative to artificial neural networks (ANNs) for targeted applications. Multi-core implementation of SNNs has been built to achieve a resource-efficient design. However, managing the spike traffic congestion while routing the spikes between different cores requires a performance-resource tradeoff to avoid any packet loss. This paper presents a novel router architecture servicing ongoing packets in a 2-D mesh network while guaranteeing no packet drop. Here, the packets are distributed across different paths to reduce spike traffic. The proposed router suitable for a 16 × 16 network occupies an area of 0.001mm2 in 28nm CMOS technology, while consuming 75 fJ/transmission.

KW - Brain-inspired computing

KW - artificial neural network

KW - neuromorphic computing

KW - spiking neural network

UR - http://www.scopus.com/inward/record.url?scp=85167675408&partnerID=8YFLogxK

U2 - 10.1109/ISCAS46773.2023.10182040

DO - 10.1109/ISCAS46773.2023.10182040

M3 - Article in proceedings

SN - 978-1-6654-5109-3

SP - 1

EP - 5

BT - ISCAS 2023 - 56th IEEE International Symposium on Circuits and Systems, Proceedings

PB - IEEE

T2 - IEEE International Symposium on Circuits and Systems

Y2 - 21 May 2023 through 25 May 2023

ER -

Hardware Implementation of a Resource-Efficient Router for Multi-Core Spiking Neural Networks

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this