Energy-Efficient Multiply-and-Accumulate using Silicon Photonics for Deep Neural Networks

Kyle Shiflett; Avinash Karanth; Ahmed Louri; Razvan Bunescu

doi:10.1109/IPC47351.2020.9252509

Energy-Efficient Multiply-and-Accumulate using Silicon Photonics for Deep Neural Networks

Kyle Shiflett, Avinash Karanth, Ahmed Louri, Razvan Bunescu

Electrical and Computer Engineering

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

1 Scopus citations

Abstract

We propose two optical hybrid matrix multipliers for deep neural networks. Our results indicate our all-optical design achieved the best performance in energy efficiency and latency, with an energy-delay product reduction of 33.1% and 76.4% for conservative and aggressive estimates, respectively.

Original language	English (US)
Title of host publication	2020 IEEE Photonics Conference, IPC 2020 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781728158914
DOIs	https://doi.org/10.1109/IPC47351.2020.9252509
State	Published - Sep 2020
Event	2020 IEEE Photonics Conference, IPC 2020 - Virtual, Vancouver, Canada Duration: Sep 28 2020 → Oct 1 2020

Publication series

Name	2020 IEEE Photonics Conference, IPC 2020 - Proceedings

Conference

Conference	2020 IEEE Photonics Conference, IPC 2020
Country/Territory	Canada
City	Virtual, Vancouver
Period	9/28/20 → 10/1/20

ASJC Scopus subject areas

Artificial Intelligence
Computer Networks and Communications
Signal Processing
Electrical and Electronic Engineering
Electronic, Optical and Magnetic Materials
Instrumentation
Atomic and Molecular Physics, and Optics

Access to Document

10.1109/IPC47351.2020.9252509

Cite this

Shiflett, K., Karanth, A., Louri, A., & Bunescu, R. (2020). Energy-Efficient Multiply-and-Accumulate using Silicon Photonics for Deep Neural Networks. In 2020 IEEE Photonics Conference, IPC 2020 - Proceedings Article 9252509 (2020 IEEE Photonics Conference, IPC 2020 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IPC47351.2020.9252509

Energy-Efficient Multiply-and-Accumulate using Silicon Photonics for Deep Neural Networks. / Shiflett, Kyle; Karanth, Avinash; Louri, Ahmed et al.
2020 IEEE Photonics Conference, IPC 2020 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2020. 9252509 (2020 IEEE Photonics Conference, IPC 2020 - Proceedings).

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

Shiflett, K, Karanth, A, Louri, A & Bunescu, R 2020, Energy-Efficient Multiply-and-Accumulate using Silicon Photonics for Deep Neural Networks. in 2020 IEEE Photonics Conference, IPC 2020 - Proceedings., 9252509, 2020 IEEE Photonics Conference, IPC 2020 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 2020 IEEE Photonics Conference, IPC 2020, Virtual, Vancouver, Canada, 9/28/20. https://doi.org/10.1109/IPC47351.2020.9252509

@inproceedings{89b32b24aab94840817bf1239a5be8a9,

title = "Energy-Efficient Multiply-and-Accumulate using Silicon Photonics for Deep Neural Networks",

abstract = "We propose two optical hybrid matrix multipliers for deep neural networks. Our results indicate our all-optical design achieved the best performance in energy efficiency and latency, with an energy-delay product reduction of 33.1% and 76.4% for conservative and aggressive estimates, respectively.",

author = "Kyle Shiflett and Avinash Karanth and Ahmed Louri and Razvan Bunescu",

note = "Funding Information: ACKNOWLEDGMENT This research was partially supported by NSF grants CCF-1513606, CCF-1703013, CCF-1901192, CCF-1513923, CCF-1547034, CCF-1547035, CCF-1547036, CCF-1702980, and CCF-1901165. Publisher Copyright: {\textcopyright} 2020 IEEE.; 2020 IEEE Photonics Conference, IPC 2020 ; Conference date: 28-09-2020 Through 01-10-2020",

year = "2020",

month = sep,

doi = "10.1109/IPC47351.2020.9252509",

language = "English (US)",

series = "2020 IEEE Photonics Conference, IPC 2020 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2020 IEEE Photonics Conference, IPC 2020 - Proceedings",

}

TY - GEN

T1 - Energy-Efficient Multiply-and-Accumulate using Silicon Photonics for Deep Neural Networks

AU - Shiflett, Kyle

AU - Karanth, Avinash

AU - Louri, Ahmed

AU - Bunescu, Razvan

N1 - Funding Information: ACKNOWLEDGMENT This research was partially supported by NSF grants CCF-1513606, CCF-1703013, CCF-1901192, CCF-1513923, CCF-1547034, CCF-1547035, CCF-1547036, CCF-1702980, and CCF-1901165. Publisher Copyright: © 2020 IEEE.

PY - 2020/9

Y1 - 2020/9

N2 - We propose two optical hybrid matrix multipliers for deep neural networks. Our results indicate our all-optical design achieved the best performance in energy efficiency and latency, with an energy-delay product reduction of 33.1% and 76.4% for conservative and aggressive estimates, respectively.

AB - We propose two optical hybrid matrix multipliers for deep neural networks. Our results indicate our all-optical design achieved the best performance in energy efficiency and latency, with an energy-delay product reduction of 33.1% and 76.4% for conservative and aggressive estimates, respectively.

UR - http://www.scopus.com/inward/record.url?scp=85097890279&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85097890279&partnerID=8YFLogxK

U2 - 10.1109/IPC47351.2020.9252509

DO - 10.1109/IPC47351.2020.9252509

M3 - Conference contribution

AN - SCOPUS:85097890279

T3 - 2020 IEEE Photonics Conference, IPC 2020 - Proceedings

BT - 2020 IEEE Photonics Conference, IPC 2020 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2020 IEEE Photonics Conference, IPC 2020

Y2 - 28 September 2020 through 1 October 2020

ER -

Energy-Efficient Multiply-and-Accumulate using Silicon Photonics for Deep Neural Networks

Abstract

Publication series

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this