NVMLearn

A simulation platform for non-volatile-memory-based deep learning hardware

Darsen Lu, Fu Xiang Liang, Yi Ci Wang, Huai Kuan Zeng

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

Abstract

Hardware implementation of deep machine learning using the convolutional neural network has been successfully demonstrated using array architecture with non-volatile storage elements such as floating-gate MOS transistor, resistive memory, phase change memory, etc. We present a new simulation platform, NVMLearn, to aid the design, verification, and system-level power and performance estimation for such architecture. Physical characteristics of memory devices are modeled using Verilog-A compact models, which can be easily simulated in SPICE to obtain the device programming, erasure, and read behavior. On the system level, NVMLearn simulates the training of the entire convolutional network based on any non-volatile memory device type.

Original languageEnglish
Title of host publicationProceedings of the 2017 IEEE International Conference on Applied System Innovation
Subtitle of host publicationApplied System Innovation for Modern Technology, ICASI 2017
EditorsTeen-Hang Meen, Artde Donald Kin-Tak Lam, Stephen D. Prior
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages66-69
Number of pages4
ISBN (Electronic)9781509048977
DOIs
Publication statusPublished - 2017 Jul 21
Event2017 IEEE International Conference on Applied System Innovation, ICASI 2017 - Sapporo, Japan
Duration: 2017 May 132017 May 17

Publication series

NameProceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017

Other

Other2017 IEEE International Conference on Applied System Innovation, ICASI 2017
CountryJapan
CitySapporo
Period17-05-1317-05-17

Fingerprint

Computer hardware
learning
hardware
platforms
Learning
Data storage equipment
Equipment and Supplies
Nonvolatile storage
Phase change memory
Computer hardware description languages
simulation
MOSFET devices
SPICE
machine learning
Learning systems
programming
floating
Neural networks
education
transistors

All Science Journal Classification (ASJC) codes

  • Computer Networks and Communications
  • Computer Science Applications
  • Hardware and Architecture
  • Safety, Risk, Reliability and Quality
  • Mechanical Engineering
  • Media Technology
  • Health Informatics
  • Instrumentation

Cite this

Lu, D., Liang, F. X., Wang, Y. C., & Zeng, H. K. (2017). NVMLearn: A simulation platform for non-volatile-memory-based deep learning hardware. In T-H. Meen, A. D. K-T. Lam, & S. D. Prior (Eds.), Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017 (pp. 66-69). [7988347] (Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASI.2017.7988347
Lu, Darsen ; Liang, Fu Xiang ; Wang, Yi Ci ; Zeng, Huai Kuan. / NVMLearn : A simulation platform for non-volatile-memory-based deep learning hardware. Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017. editor / Teen-Hang Meen ; Artde Donald Kin-Tak Lam ; Stephen D. Prior. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 66-69 (Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017).
@inproceedings{09e6e7f5b1154663baf7632c19222df1,
title = "NVMLearn: A simulation platform for non-volatile-memory-based deep learning hardware",
abstract = "Hardware implementation of deep machine learning using the convolutional neural network has been successfully demonstrated using array architecture with non-volatile storage elements such as floating-gate MOS transistor, resistive memory, phase change memory, etc. We present a new simulation platform, NVMLearn, to aid the design, verification, and system-level power and performance estimation for such architecture. Physical characteristics of memory devices are modeled using Verilog-A compact models, which can be easily simulated in SPICE to obtain the device programming, erasure, and read behavior. On the system level, NVMLearn simulates the training of the entire convolutional network based on any non-volatile memory device type.",
author = "Darsen Lu and Liang, {Fu Xiang} and Wang, {Yi Ci} and Zeng, {Huai Kuan}",
year = "2017",
month = "7",
day = "21",
doi = "10.1109/ICASI.2017.7988347",
language = "English",
series = "Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "66--69",
editor = "Teen-Hang Meen and Lam, {Artde Donald Kin-Tak} and Prior, {Stephen D.}",
booktitle = "Proceedings of the 2017 IEEE International Conference on Applied System Innovation",
address = "United States",

}

Lu, D, Liang, FX, Wang, YC & Zeng, HK 2017, NVMLearn: A simulation platform for non-volatile-memory-based deep learning hardware. in T-H Meen, ADK-T Lam & SD Prior (eds), Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017., 7988347, Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017, Institute of Electrical and Electronics Engineers Inc., pp. 66-69, 2017 IEEE International Conference on Applied System Innovation, ICASI 2017, Sapporo, Japan, 17-05-13. https://doi.org/10.1109/ICASI.2017.7988347

NVMLearn : A simulation platform for non-volatile-memory-based deep learning hardware. / Lu, Darsen; Liang, Fu Xiang; Wang, Yi Ci; Zeng, Huai Kuan.

Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017. ed. / Teen-Hang Meen; Artde Donald Kin-Tak Lam; Stephen D. Prior. Institute of Electrical and Electronics Engineers Inc., 2017. p. 66-69 7988347 (Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017).

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

TY - GEN

T1 - NVMLearn

T2 - A simulation platform for non-volatile-memory-based deep learning hardware

AU - Lu, Darsen

AU - Liang, Fu Xiang

AU - Wang, Yi Ci

AU - Zeng, Huai Kuan

PY - 2017/7/21

Y1 - 2017/7/21

N2 - Hardware implementation of deep machine learning using the convolutional neural network has been successfully demonstrated using array architecture with non-volatile storage elements such as floating-gate MOS transistor, resistive memory, phase change memory, etc. We present a new simulation platform, NVMLearn, to aid the design, verification, and system-level power and performance estimation for such architecture. Physical characteristics of memory devices are modeled using Verilog-A compact models, which can be easily simulated in SPICE to obtain the device programming, erasure, and read behavior. On the system level, NVMLearn simulates the training of the entire convolutional network based on any non-volatile memory device type.

AB - Hardware implementation of deep machine learning using the convolutional neural network has been successfully demonstrated using array architecture with non-volatile storage elements such as floating-gate MOS transistor, resistive memory, phase change memory, etc. We present a new simulation platform, NVMLearn, to aid the design, verification, and system-level power and performance estimation for such architecture. Physical characteristics of memory devices are modeled using Verilog-A compact models, which can be easily simulated in SPICE to obtain the device programming, erasure, and read behavior. On the system level, NVMLearn simulates the training of the entire convolutional network based on any non-volatile memory device type.

UR - http://www.scopus.com/inward/record.url?scp=85028540400&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85028540400&partnerID=8YFLogxK

U2 - 10.1109/ICASI.2017.7988347

DO - 10.1109/ICASI.2017.7988347

M3 - Conference contribution

T3 - Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017

SP - 66

EP - 69

BT - Proceedings of the 2017 IEEE International Conference on Applied System Innovation

A2 - Meen, Teen-Hang

A2 - Lam, Artde Donald Kin-Tak

A2 - Prior, Stephen D.

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Lu D, Liang FX, Wang YC, Zeng HK. NVMLearn: A simulation platform for non-volatile-memory-based deep learning hardware. In Meen T-H, Lam ADK-T, Prior SD, editors, Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 66-69. 7988347. (Proceedings of the 2017 IEEE International Conference on Applied System Innovation: Applied System Innovation for Modern Technology, ICASI 2017). https://doi.org/10.1109/ICASI.2017.7988347