Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture

Maarten Molendijk; Kanishkan Vadivel; Federico Corradi; Gert-Jan van Schaik; Amirreza Yousefzadeh; Henk Corporaal

doi:10.13052/rp-9788770227902

Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture

Maarten Molendijk
, Kanishkan Vadivel
, Federico Corradi
, Gert-Jan van Schaik
, Amirreza Yousefzadeh
, Henk Corporaal

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Professional

289 Downloads (Pure)

Abstract

This short article explains why the Epiphany architecture is a proper refer-
ence for digital large-scale neuromorphic design. We compare the Epiphany
architecture with several modern digital neuromorphic processors. We show
the result of mapping the binary LeNet-5 neural network into few modern
neuromorphic architectures and demonstrate the efficient use of memory in
Epiphany. Finally, we show the results of our benchmarking experiments
with Epiphany and propose a few suggestions to improve the architecture
for neuromorphic applications. Epiphany can update a neuron on average in
120ns which is enough for many real-time neuromorphic applications.

Original language	English
Title of host publication	Industrial Artificial Intelligence Technologies and Applications
Editors	Ovidiu Vermesan
Publisher	River Publishers
Chapter	2
Pages	21-34
Number of pages	14
ISBN (Electronic)	9788770227902
ISBN (Print)	9788770227919
DOIs	https://doi.org/10.13052/rp-9788770227902
Publication status	Published - Jun 2022

Publication series

Name	River Publishers Series in Communications and Networking

Access to Document

10.13052/rp-9788770227902

pdfFinal published version, 404 KB

Cite this

Molendijk, M., Vadivel, K., Corradi, F., van Schaik, G.-J., Yousefzadeh, A., & Corporaal, H. (2022). Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture. In O. Vermesan (Ed.), Industrial Artificial Intelligence Technologies and Applications (pp. 21-34). (River Publishers Series in Communications and Networking ). River Publishers. https://doi.org/10.13052/rp-9788770227902

@inbook{171a580674ab483fae19871c5142ce32,

title = "Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture",

abstract = "This short article explains why the Epiphany architecture is a proper refer- ence for digital large-scale neuromorphic design. We compare the Epiphany architecture with several modern digital neuromorphic processors. We show the result of mapping the binary LeNet-5 neural network into few modern neuromorphic architectures and demonstrate the efficient use of memory in Epiphany. Finally, we show the results of our benchmarking experiments with Epiphany and propose a few suggestions to improve the architecture for neuromorphic applications. Epiphany can update a neuron on average in 120ns which is enough for many real-time neuromorphic applications.",

author = "Maarten Molendijk and Kanishkan Vadivel and Federico Corradi and \{van Schaik\}, Gert-Jan and Amirreza Yousefzadeh and Henk Corporaal",

year = "2022",

month = jun,

doi = "10.13052/rp-9788770227902",

language = "English",

isbn = "9788770227919",

series = "River Publishers Series in Communications and Networking ",

publisher = "River Publishers",

pages = "21--34",

editor = "Ovidiu Vermesan",

booktitle = "Industrial Artificial Intelligence Technologies and Applications",

address = "Denmark",

}

Molendijk, M , Vadivel, K , Corradi, F, van Schaik, G-J, Yousefzadeh, A & Corporaal, H 2022, Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture. in O Vermesan (ed.), Industrial Artificial Intelligence Technologies and Applications . River Publishers Series in Communications and Networking , River Publishers, pp. 21-34. https://doi.org/10.13052/rp-9788770227902

Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture. / Molendijk, Maarten ; Vadivel, Kanishkan ; Corradi, Federico et al.
Industrial Artificial Intelligence Technologies and Applications . ed. / Ovidiu Vermesan. River Publishers, 2022. p. 21-34 (River Publishers Series in Communications and Networking ).

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Professional

TY - CHAP

T1 - Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture

AU - Molendijk, Maarten

AU - Vadivel, Kanishkan

AU - Corradi, Federico

AU - van Schaik, Gert-Jan

AU - Yousefzadeh, Amirreza

AU - Corporaal, Henk

PY - 2022/6

Y1 - 2022/6

N2 - This short article explains why the Epiphany architecture is a proper refer- ence for digital large-scale neuromorphic design. We compare the Epiphany architecture with several modern digital neuromorphic processors. We show the result of mapping the binary LeNet-5 neural network into few modern neuromorphic architectures and demonstrate the efficient use of memory in Epiphany. Finally, we show the results of our benchmarking experiments with Epiphany and propose a few suggestions to improve the architecture for neuromorphic applications. Epiphany can update a neuron on average in 120ns which is enough for many real-time neuromorphic applications.

AB - This short article explains why the Epiphany architecture is a proper refer- ence for digital large-scale neuromorphic design. We compare the Epiphany architecture with several modern digital neuromorphic processors. We show the result of mapping the binary LeNet-5 neural network into few modern neuromorphic architectures and demonstrate the efficient use of memory in Epiphany. Finally, we show the results of our benchmarking experiments with Epiphany and propose a few suggestions to improve the architecture for neuromorphic applications. Epiphany can update a neuron on average in 120ns which is enough for many real-time neuromorphic applications.

U2 - 10.13052/rp-9788770227902

DO - 10.13052/rp-9788770227902

M3 - Chapter

SN - 9788770227919

T3 - River Publishers Series in Communications and Networking

SP - 21

EP - 34

BT - Industrial Artificial Intelligence Technologies and Applications

A2 - Vermesan, Ovidiu

PB - River Publishers

ER -

Molendijk M , Vadivel K , Corradi F, van Schaik GJ, Yousefzadeh A, Corporaal H. Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture. In Vermesan O, editor, Industrial Artificial Intelligence Technologies and Applications . River Publishers. 2022. p. 21-34. (River Publishers Series in Communications and Networking ). doi: 10.13052/rp-9788770227902

Benchmarking the Epiphany Processor as a Reference Neuromorphic Architecture

Abstract

Publication series

Access to Document

Fingerprint

Cite this