Abstract

Neuromorphic computing systems built with resistive random access memory (RRAM) are attractive solutions for implementing deep learning on-chip. In this work, the single event and cumulative radiation damage susceptibility of a 1-transistor-1-resistor (1T1R) synaptic array architecture is investigated for the inference stage of a neuromorphic system. A physics-based SPICE RRAM model is developed to simulate the analog switching behavior of HfOx-based RRAM devices. SPICE simulations using the RRAM model are performed to capture the effects of transient heavy ion strikes and experimental results obtained on GeSe based RRAM devices are used to model the effects of cumulative radiation dose. Error patterns are fed into the software simulation of a multilayer perceptron (MLP), a representative artificial neural network for MNIST handwritten digit recognition. Radiation effects on the inference accuracy are analyzed. The results of the analysis indicate that the RRAM-based neuromorphic computing system is highly resistant to transient single event effects (SEEs) due to the low cross-section. The system is also shown to be tolerant to multi Mrad levels of total ionizing and displacement damage dose.

Original languageEnglish (US)
JournalIEEE Transactions on Nuclear Science
DOIs
StateAccepted/In press - Jan 1 2018

Fingerprint

Radiation effects
random access memory
radiation effects
inference
Data storage equipment
evaluation
SPICE
cumulative damage
dosage
self organizing systems
digits
Radiation damage
Multilayer neural networks
Heavy ions
radiation damage
resistors
Resistors
learning
Dosimetry
heavy ions

Keywords

  • analog RRAM
  • Hafnium compounds
  • Ions
  • multilayer perceptron
  • neuromorphic computing
  • Neuromorphics
  • Resistance
  • single-event upset
  • SPICE
  • Switches
  • total ionizing dose
  • Transient analysis

ASJC Scopus subject areas

  • Nuclear and High Energy Physics
  • Nuclear Energy and Engineering
  • Electrical and Electronic Engineering

Cite this

Evaluation of Radiation Effects in RRAM based Neuromorphic Computing System for Inference. / Ye, Zhilu; Liu, Rui; Taggart, Jennifer; Barnaby, Hugh; Yu, Shimeng.

In: IEEE Transactions on Nuclear Science, 01.01.2018.

Research output: Contribution to journalArticle

@article{ba0a7063a05d4cbe8c0990cd8b1262bf,
title = "Evaluation of Radiation Effects in RRAM based Neuromorphic Computing System for Inference",
abstract = "Neuromorphic computing systems built with resistive random access memory (RRAM) are attractive solutions for implementing deep learning on-chip. In this work, the single event and cumulative radiation damage susceptibility of a 1-transistor-1-resistor (1T1R) synaptic array architecture is investigated for the inference stage of a neuromorphic system. A physics-based SPICE RRAM model is developed to simulate the analog switching behavior of HfOx-based RRAM devices. SPICE simulations using the RRAM model are performed to capture the effects of transient heavy ion strikes and experimental results obtained on GeSe based RRAM devices are used to model the effects of cumulative radiation dose. Error patterns are fed into the software simulation of a multilayer perceptron (MLP), a representative artificial neural network for MNIST handwritten digit recognition. Radiation effects on the inference accuracy are analyzed. The results of the analysis indicate that the RRAM-based neuromorphic computing system is highly resistant to transient single event effects (SEEs) due to the low cross-section. The system is also shown to be tolerant to multi Mrad levels of total ionizing and displacement damage dose.",
keywords = "analog RRAM, Hafnium compounds, Ions, multilayer perceptron, neuromorphic computing, Neuromorphics, Resistance, single-event upset, SPICE, Switches, total ionizing dose, Transient analysis",
author = "Zhilu Ye and Rui Liu and Jennifer Taggart and Hugh Barnaby and Shimeng Yu",
year = "2018",
month = "1",
day = "1",
doi = "10.1109/TNS.2018.2886793",
language = "English (US)",
journal = "IEEE Transactions on Nuclear Science",
issn = "0018-9499",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Evaluation of Radiation Effects in RRAM based Neuromorphic Computing System for Inference

AU - Ye, Zhilu

AU - Liu, Rui

AU - Taggart, Jennifer

AU - Barnaby, Hugh

AU - Yu, Shimeng

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Neuromorphic computing systems built with resistive random access memory (RRAM) are attractive solutions for implementing deep learning on-chip. In this work, the single event and cumulative radiation damage susceptibility of a 1-transistor-1-resistor (1T1R) synaptic array architecture is investigated for the inference stage of a neuromorphic system. A physics-based SPICE RRAM model is developed to simulate the analog switching behavior of HfOx-based RRAM devices. SPICE simulations using the RRAM model are performed to capture the effects of transient heavy ion strikes and experimental results obtained on GeSe based RRAM devices are used to model the effects of cumulative radiation dose. Error patterns are fed into the software simulation of a multilayer perceptron (MLP), a representative artificial neural network for MNIST handwritten digit recognition. Radiation effects on the inference accuracy are analyzed. The results of the analysis indicate that the RRAM-based neuromorphic computing system is highly resistant to transient single event effects (SEEs) due to the low cross-section. The system is also shown to be tolerant to multi Mrad levels of total ionizing and displacement damage dose.

AB - Neuromorphic computing systems built with resistive random access memory (RRAM) are attractive solutions for implementing deep learning on-chip. In this work, the single event and cumulative radiation damage susceptibility of a 1-transistor-1-resistor (1T1R) synaptic array architecture is investigated for the inference stage of a neuromorphic system. A physics-based SPICE RRAM model is developed to simulate the analog switching behavior of HfOx-based RRAM devices. SPICE simulations using the RRAM model are performed to capture the effects of transient heavy ion strikes and experimental results obtained on GeSe based RRAM devices are used to model the effects of cumulative radiation dose. Error patterns are fed into the software simulation of a multilayer perceptron (MLP), a representative artificial neural network for MNIST handwritten digit recognition. Radiation effects on the inference accuracy are analyzed. The results of the analysis indicate that the RRAM-based neuromorphic computing system is highly resistant to transient single event effects (SEEs) due to the low cross-section. The system is also shown to be tolerant to multi Mrad levels of total ionizing and displacement damage dose.

KW - analog RRAM

KW - Hafnium compounds

KW - Ions

KW - multilayer perceptron

KW - neuromorphic computing

KW - Neuromorphics

KW - Resistance

KW - single-event upset

KW - SPICE

KW - Switches

KW - total ionizing dose

KW - Transient analysis

UR - http://www.scopus.com/inward/record.url?scp=85058899960&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058899960&partnerID=8YFLogxK

U2 - 10.1109/TNS.2018.2886793

DO - 10.1109/TNS.2018.2886793

M3 - Article

JO - IEEE Transactions on Nuclear Science

JF - IEEE Transactions on Nuclear Science

SN - 0018-9499

ER -