Domain Wall Memory based Convolutional Neural Networks for Bit-width Extendability and Energy-Efficiency

Jinil Chung, Jongsun Park, Swaroop Ghosh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

In the hardware implementation of deep learning algorithms such as Convolutional Neural Networks (CNNs), vector-vector multiplications and memories for storing parameters take a significant portion of area and power consumption. In this paper, we propose a Domain Wall Memory (DWM) based design of CNN convolutional layer. In the proposed design, the resistive cell sensing mechanism is efficiently exploited to design a low-cost DWM-based cell arrays for storing parameters. The unique serial access mechanism and small footprint of DWM are also used to reduce the area and power cost of the input registers for aligning inputs. Contrary to the conventional implementation using Memristor-Based Crossbar (MBC), the bit-width of the proposed CNN convolutional layer is extendable for high resolution classifications and training. Simulation results using 65 nm CMOS process show that the proposed design archives 34% of energy savings compared to the conventional MBC based design approach.

Original languageEnglish (US)
Title of host publicationISLPED 2016 - Proceedings of the 2016 International Symposium on Low Power Electronics and Design
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages332-337
Number of pages6
ISBN (Electronic)9781450341851
DOIs
StatePublished - Aug 8 2016
Event21st IEEE/ACM International Symposium on Low Power Electronics and Design, ISLPED 2016 - San Francisco, United States
Duration: Aug 8 2016Aug 10 2016

Publication series

NameProceedings of the International Symposium on Low Power Electronics and Design
ISSN (Print)1533-4678

Other

Other21st IEEE/ACM International Symposium on Low Power Electronics and Design, ISLPED 2016
CountryUnited States
CitySan Francisco
Period8/8/168/10/16

Fingerprint

Domain walls
Energy efficiency
Neural networks
Data storage equipment
Memristors
Learning algorithms
Costs
Energy conservation
Electric power utilization
Hardware

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Chung, J., Park, J., & Ghosh, S. (2016). Domain Wall Memory based Convolutional Neural Networks for Bit-width Extendability and Energy-Efficiency. In ISLPED 2016 - Proceedings of the 2016 International Symposium on Low Power Electronics and Design (pp. 332-337). (Proceedings of the International Symposium on Low Power Electronics and Design). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1145/2934583.2934602
Chung, Jinil ; Park, Jongsun ; Ghosh, Swaroop. / Domain Wall Memory based Convolutional Neural Networks for Bit-width Extendability and Energy-Efficiency. ISLPED 2016 - Proceedings of the 2016 International Symposium on Low Power Electronics and Design. Institute of Electrical and Electronics Engineers Inc., 2016. pp. 332-337 (Proceedings of the International Symposium on Low Power Electronics and Design).
@inproceedings{deb5e37927c04de29b216a92be161d13,
title = "Domain Wall Memory based Convolutional Neural Networks for Bit-width Extendability and Energy-Efficiency",
abstract = "In the hardware implementation of deep learning algorithms such as Convolutional Neural Networks (CNNs), vector-vector multiplications and memories for storing parameters take a significant portion of area and power consumption. In this paper, we propose a Domain Wall Memory (DWM) based design of CNN convolutional layer. In the proposed design, the resistive cell sensing mechanism is efficiently exploited to design a low-cost DWM-based cell arrays for storing parameters. The unique serial access mechanism and small footprint of DWM are also used to reduce the area and power cost of the input registers for aligning inputs. Contrary to the conventional implementation using Memristor-Based Crossbar (MBC), the bit-width of the proposed CNN convolutional layer is extendable for high resolution classifications and training. Simulation results using 65 nm CMOS process show that the proposed design archives 34{\%} of energy savings compared to the conventional MBC based design approach.",
author = "Jinil Chung and Jongsun Park and Swaroop Ghosh",
year = "2016",
month = "8",
day = "8",
doi = "10.1145/2934583.2934602",
language = "English (US)",
series = "Proceedings of the International Symposium on Low Power Electronics and Design",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "332--337",
booktitle = "ISLPED 2016 - Proceedings of the 2016 International Symposium on Low Power Electronics and Design",
address = "United States",

}

Chung, J, Park, J & Ghosh, S 2016, Domain Wall Memory based Convolutional Neural Networks for Bit-width Extendability and Energy-Efficiency. in ISLPED 2016 - Proceedings of the 2016 International Symposium on Low Power Electronics and Design. Proceedings of the International Symposium on Low Power Electronics and Design, Institute of Electrical and Electronics Engineers Inc., pp. 332-337, 21st IEEE/ACM International Symposium on Low Power Electronics and Design, ISLPED 2016, San Francisco, United States, 8/8/16. https://doi.org/10.1145/2934583.2934602

Domain Wall Memory based Convolutional Neural Networks for Bit-width Extendability and Energy-Efficiency. / Chung, Jinil; Park, Jongsun; Ghosh, Swaroop.

ISLPED 2016 - Proceedings of the 2016 International Symposium on Low Power Electronics and Design. Institute of Electrical and Electronics Engineers Inc., 2016. p. 332-337 (Proceedings of the International Symposium on Low Power Electronics and Design).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Domain Wall Memory based Convolutional Neural Networks for Bit-width Extendability and Energy-Efficiency

AU - Chung, Jinil

AU - Park, Jongsun

AU - Ghosh, Swaroop

PY - 2016/8/8

Y1 - 2016/8/8

N2 - In the hardware implementation of deep learning algorithms such as Convolutional Neural Networks (CNNs), vector-vector multiplications and memories for storing parameters take a significant portion of area and power consumption. In this paper, we propose a Domain Wall Memory (DWM) based design of CNN convolutional layer. In the proposed design, the resistive cell sensing mechanism is efficiently exploited to design a low-cost DWM-based cell arrays for storing parameters. The unique serial access mechanism and small footprint of DWM are also used to reduce the area and power cost of the input registers for aligning inputs. Contrary to the conventional implementation using Memristor-Based Crossbar (MBC), the bit-width of the proposed CNN convolutional layer is extendable for high resolution classifications and training. Simulation results using 65 nm CMOS process show that the proposed design archives 34% of energy savings compared to the conventional MBC based design approach.

AB - In the hardware implementation of deep learning algorithms such as Convolutional Neural Networks (CNNs), vector-vector multiplications and memories for storing parameters take a significant portion of area and power consumption. In this paper, we propose a Domain Wall Memory (DWM) based design of CNN convolutional layer. In the proposed design, the resistive cell sensing mechanism is efficiently exploited to design a low-cost DWM-based cell arrays for storing parameters. The unique serial access mechanism and small footprint of DWM are also used to reduce the area and power cost of the input registers for aligning inputs. Contrary to the conventional implementation using Memristor-Based Crossbar (MBC), the bit-width of the proposed CNN convolutional layer is extendable for high resolution classifications and training. Simulation results using 65 nm CMOS process show that the proposed design archives 34% of energy savings compared to the conventional MBC based design approach.

UR - http://www.scopus.com/inward/record.url?scp=84988007983&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84988007983&partnerID=8YFLogxK

U2 - 10.1145/2934583.2934602

DO - 10.1145/2934583.2934602

M3 - Conference contribution

AN - SCOPUS:84988007983

T3 - Proceedings of the International Symposium on Low Power Electronics and Design

SP - 332

EP - 337

BT - ISLPED 2016 - Proceedings of the 2016 International Symposium on Low Power Electronics and Design

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Chung J, Park J, Ghosh S. Domain Wall Memory based Convolutional Neural Networks for Bit-width Extendability and Energy-Efficiency. In ISLPED 2016 - Proceedings of the 2016 International Symposium on Low Power Electronics and Design. Institute of Electrical and Electronics Engineers Inc. 2016. p. 332-337. (Proceedings of the International Symposium on Low Power Electronics and Design). https://doi.org/10.1145/2934583.2934602