Parallel programming of resistive cross-point array for synaptic plasticity

Zihan Xu, Abinash Mohanty, Pai Yu Chen, Deepak Kadetotad, Binbin Lin, Jieping Ye, Sarma Vrudhula, Shimeng Yu, Jae-sun Seo, Yu Cao

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

10 Citations (Scopus)

Abstract

This paper proposes a parallel programming scheme for the cross-point array with resistive random access memory (RRAM). Synaptic plasticity in unsupervised learning is realized by tuning the conductance of each RRAM cell. Inspired by the spike-timing-dependent-plasticity (STDP), the programming strength is encoded into the spike firing rate (i.e., pulse frequency) and the overlap time (i.e., duty cycle) of the pre-synaptic node and post-synaptic node, and simultaneously applied to all RRAM cells in the cross-point array. Such an approach achieves parallel programming of the entire RRAM array, only requiring local information from pre-synaptic and post-synaptic nodes to each RRAM cell. As demonstrated by digital peripheral circuits implemented in 65nm CMOS, the programming time of a 40kb RRAM array is 84 ns, indicating 900X speedup as compared to state-ofthe-art software approach of sparse coding in image feature extraction.

Original languageEnglish (US)
Title of host publicationProcedia Computer Science
PublisherElsevier
Pages126-133
Number of pages8
Volume41
DOIs
StatePublished - 2014
Event5th Annual International Conference on Biologically Inspired Cognitive Architectures, BICA 2014 - Cambridge, United States
Duration: Nov 7 2014Nov 9 2014

Other

Other5th Annual International Conference on Biologically Inspired Cognitive Architectures, BICA 2014
CountryUnited States
CityCambridge
Period11/7/1411/9/14

Fingerprint

Parallel programming
Plasticity
Data storage equipment
Computer peripheral equipment
Unsupervised learning
Computer programming
Feature extraction
Tuning
Networks (circuits)

Keywords

  • Dictionary learning
  • Parallel programming
  • Resistive cross-point array
  • Synaptic plasticity

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Xu, Z., Mohanty, A., Chen, P. Y., Kadetotad, D., Lin, B., Ye, J., ... Cao, Y. (2014). Parallel programming of resistive cross-point array for synaptic plasticity. In Procedia Computer Science (Vol. 41, pp. 126-133). Elsevier. https://doi.org/10.1016/j.procs.2014.11.094

Parallel programming of resistive cross-point array for synaptic plasticity. / Xu, Zihan; Mohanty, Abinash; Chen, Pai Yu; Kadetotad, Deepak; Lin, Binbin; Ye, Jieping; Vrudhula, Sarma; Yu, Shimeng; Seo, Jae-sun; Cao, Yu.

Procedia Computer Science. Vol. 41 Elsevier, 2014. p. 126-133.

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

Xu, Z, Mohanty, A, Chen, PY, Kadetotad, D, Lin, B, Ye, J, Vrudhula, S, Yu, S, Seo, J & Cao, Y 2014, Parallel programming of resistive cross-point array for synaptic plasticity. in Procedia Computer Science. vol. 41, Elsevier, pp. 126-133, 5th Annual International Conference on Biologically Inspired Cognitive Architectures, BICA 2014, Cambridge, United States, 11/7/14. https://doi.org/10.1016/j.procs.2014.11.094
Xu Z, Mohanty A, Chen PY, Kadetotad D, Lin B, Ye J et al. Parallel programming of resistive cross-point array for synaptic plasticity. In Procedia Computer Science. Vol. 41. Elsevier. 2014. p. 126-133 https://doi.org/10.1016/j.procs.2014.11.094
Xu, Zihan ; Mohanty, Abinash ; Chen, Pai Yu ; Kadetotad, Deepak ; Lin, Binbin ; Ye, Jieping ; Vrudhula, Sarma ; Yu, Shimeng ; Seo, Jae-sun ; Cao, Yu. / Parallel programming of resistive cross-point array for synaptic plasticity. Procedia Computer Science. Vol. 41 Elsevier, 2014. pp. 126-133
@inproceedings{c7c13d07f0f549648616b30fea13205d,
title = "Parallel programming of resistive cross-point array for synaptic plasticity",
abstract = "This paper proposes a parallel programming scheme for the cross-point array with resistive random access memory (RRAM). Synaptic plasticity in unsupervised learning is realized by tuning the conductance of each RRAM cell. Inspired by the spike-timing-dependent-plasticity (STDP), the programming strength is encoded into the spike firing rate (i.e., pulse frequency) and the overlap time (i.e., duty cycle) of the pre-synaptic node and post-synaptic node, and simultaneously applied to all RRAM cells in the cross-point array. Such an approach achieves parallel programming of the entire RRAM array, only requiring local information from pre-synaptic and post-synaptic nodes to each RRAM cell. As demonstrated by digital peripheral circuits implemented in 65nm CMOS, the programming time of a 40kb RRAM array is 84 ns, indicating 900X speedup as compared to state-ofthe-art software approach of sparse coding in image feature extraction.",
keywords = "Dictionary learning, Parallel programming, Resistive cross-point array, Synaptic plasticity",
author = "Zihan Xu and Abinash Mohanty and Chen, {Pai Yu} and Deepak Kadetotad and Binbin Lin and Jieping Ye and Sarma Vrudhula and Shimeng Yu and Jae-sun Seo and Yu Cao",
year = "2014",
doi = "10.1016/j.procs.2014.11.094",
language = "English (US)",
volume = "41",
pages = "126--133",
booktitle = "Procedia Computer Science",
publisher = "Elsevier",

}

TY - GEN

T1 - Parallel programming of resistive cross-point array for synaptic plasticity

AU - Xu, Zihan

AU - Mohanty, Abinash

AU - Chen, Pai Yu

AU - Kadetotad, Deepak

AU - Lin, Binbin

AU - Ye, Jieping

AU - Vrudhula, Sarma

AU - Yu, Shimeng

AU - Seo, Jae-sun

AU - Cao, Yu

PY - 2014

Y1 - 2014

N2 - This paper proposes a parallel programming scheme for the cross-point array with resistive random access memory (RRAM). Synaptic plasticity in unsupervised learning is realized by tuning the conductance of each RRAM cell. Inspired by the spike-timing-dependent-plasticity (STDP), the programming strength is encoded into the spike firing rate (i.e., pulse frequency) and the overlap time (i.e., duty cycle) of the pre-synaptic node and post-synaptic node, and simultaneously applied to all RRAM cells in the cross-point array. Such an approach achieves parallel programming of the entire RRAM array, only requiring local information from pre-synaptic and post-synaptic nodes to each RRAM cell. As demonstrated by digital peripheral circuits implemented in 65nm CMOS, the programming time of a 40kb RRAM array is 84 ns, indicating 900X speedup as compared to state-ofthe-art software approach of sparse coding in image feature extraction.

AB - This paper proposes a parallel programming scheme for the cross-point array with resistive random access memory (RRAM). Synaptic plasticity in unsupervised learning is realized by tuning the conductance of each RRAM cell. Inspired by the spike-timing-dependent-plasticity (STDP), the programming strength is encoded into the spike firing rate (i.e., pulse frequency) and the overlap time (i.e., duty cycle) of the pre-synaptic node and post-synaptic node, and simultaneously applied to all RRAM cells in the cross-point array. Such an approach achieves parallel programming of the entire RRAM array, only requiring local information from pre-synaptic and post-synaptic nodes to each RRAM cell. As demonstrated by digital peripheral circuits implemented in 65nm CMOS, the programming time of a 40kb RRAM array is 84 ns, indicating 900X speedup as compared to state-ofthe-art software approach of sparse coding in image feature extraction.

KW - Dictionary learning

KW - Parallel programming

KW - Resistive cross-point array

KW - Synaptic plasticity

UR - http://www.scopus.com/inward/record.url?scp=84939225102&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84939225102&partnerID=8YFLogxK

U2 - 10.1016/j.procs.2014.11.094

DO - 10.1016/j.procs.2014.11.094

M3 - Conference contribution

AN - SCOPUS:84939225102

VL - 41

SP - 126

EP - 133

BT - Procedia Computer Science

PB - Elsevier

ER -