Compact Modeling and Mitigation of Parasitics in Crosspoint Accelerators of Neural Networks

In-memory computing (IMC) can accelerate data-intensive tasks, such as matrix-vector multiplication (MVM) or artificial neural networks (ANNs) inference, by means of the crosspoint memory array, allowing to reduce time and energy consumption. IMC accuracy, however, is affected by nonidealities, such...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE transactions on electron devices Ročník 71; číslo 3; s. 1 - 7
Hlavní autoři: Lepri, N., Glukhov, A., Mannocci, P., Porzani, M., Ielmini, D.
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York IEEE 01.03.2024
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:
ISSN:0018-9383, 1557-9646
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract In-memory computing (IMC) can accelerate data-intensive tasks, such as matrix-vector multiplication (MVM) or artificial neural networks (ANNs) inference, by means of the crosspoint memory array, allowing to reduce time and energy consumption. IMC accuracy, however, is affected by nonidealities, such as variability of the conductive weights or IR drop along wires due to parasitic resistances, whose impact steeply increases with the increase of array size. This work proposes a compact model to assess the impact of nonidealities for various circuital implementations, together with architectural schemes for their mitigation based on replicated arrays. The proposed mitigation techniques allow to restore the ANN accuracy from 72.7% to 94.9%, close to the software accuracy of 96.9%, in view of an increased area and energy consumption.
AbstractList In-memory computing (IMC) can accelerate data-intensive tasks, such as matrix-vector multiplication (MVM) or artificial neural networks (ANNs) inference, by means of the crosspoint memory array, allowing to reduce time and energy consumption. IMC accuracy, however, is affected by nonidealities, such as variability of the conductive weights or IR drop along wires due to parasitic resistances, whose impact steeply increases with the increase of array size. This work proposes a compact model to assess the impact of nonidealities for various circuital implementations, together with architectural schemes for their mitigation based on replicated arrays. The proposed mitigation techniques allow to restore the ANN accuracy from 72.7% to 94.9%, close to the software accuracy of 96.9%, in view of an increased area and energy consumption.
Author Ielmini, D.
Lepri, N.
Mannocci, P.
Porzani, M.
Glukhov, A.
Author_xml – sequence: 1
  givenname: N.
  orcidid: 0000-0001-7602-4058
  surname: Lepri
  fullname: Lepri, N.
  organization: Dipartimento di Elettronica, Informazione e Bioingegneria, IU.NET, Politecnico di Milano, Milan, Italy
– sequence: 2
  givenname: A.
  orcidid: 0000-0001-9362-2113
  surname: Glukhov
  fullname: Glukhov, A.
  organization: Dipartimento di Elettronica, Informazione e Bioingegneria, IU.NET, Politecnico di Milano, Milan, Italy
– sequence: 3
  givenname: P.
  orcidid: 0000-0002-0083-5804
  surname: Mannocci
  fullname: Mannocci, P.
  organization: Dipartimento di Elettronica, Informazione e Bioingegneria, IU.NET, Politecnico di Milano, Milan, Italy
– sequence: 4
  givenname: M.
  orcidid: 0009-0004-9023-6101
  surname: Porzani
  fullname: Porzani, M.
  organization: Dipartimento di Elettronica, Informazione e Bioingegneria, IU.NET, Politecnico di Milano, Milan, Italy
– sequence: 5
  givenname: D.
  orcidid: 0000-0002-1853-1614
  surname: Ielmini
  fullname: Ielmini, D.
  organization: Dipartimento di Elettronica, Informazione e Bioingegneria, IU.NET, Politecnico di Milano, Milan, Italy
BookMark eNp9kL1PwzAQxS1UJNrCzsAQiTnFsR0nHqtQPqQWGMrEYDnOpXJJ42C7Qvz3JLQDYmA63dN793S_CRq1tgWELhM8SxIsbtaL2xnBhM0o5Rgn6QkaJ2maxYIzPkLjXspjQXN6hibeb_uVM0bG6K2wu07pEK1sBY1pN5Fqq2hlgtmoYGwb2Tp6UU75XtE-Mm1UOOt9Z00bornW0IBTwTo_GJ9g71TTj_Bp3bs_R6e1ajxcHOcUvd4t1sVDvHy-fyzmy1hTykLMGKc5qUjNMlqmWOCKAa-qjJQZEEg5lBXVokoVFQmDGpepKEusMg2ElSUVdIquD3c7Zz_24IPc2r1r-0pJxIBD5Hhw4YNLDw84qGXnzE65L5lgOSCUPUI5IJRHhH2E_4loE36wBKdM81_w6hA0APCrhxHKCKff3jeAgQ
CODEN IETDAI
CitedBy_id crossref_primary_10_1021_acs_chemrev_4c00845
crossref_primary_10_1109_TED_2025_3573996
crossref_primary_10_1038_s43588_025_00854_1
crossref_primary_10_1109_TED_2024_3480898
Cites_doi 10.1016/j.mee.2019.05.004
10.1109/LED.2023.3285916
10.1109/5.726791
10.1109/TED.2021.3095433
10.1109/IEDM.2007.4419107
10.1109/ESSCIRC.2007.4430310
10.1109/iedm19573.2019.8993599
10.3389/fnins.2020.00634
10.1038/s41928-018-0092-2
10.3389/fnins.2016.00333
10.1109/TED.2021.3089995
10.1038/s41928-017-0002-z
10.1116/1.1642639
10.1109/ted.2022.3169112
10.1109/TED.2022.3141987
10.1109/IRPS48227.2022.9764486
10.1145/3316781.3317872
10.1109/ISSCC19947.2020.9062953
10.1002/advs.202105784
10.1109/JETCAS.2022.3172170
10.1109/TED.2018.2865352
ContentType Journal Article
Copyright Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024
Copyright_xml – notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024
DBID 97E
RIA
RIE
AAYXX
CITATION
7SP
8FD
L7M
DOI 10.1109/TED.2024.3360015
DatabaseName IEEE All-Society Periodicals Package (ASPP) 2005–Present
IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE Electronic Library (IEL)
CrossRef
Electronics & Communications Abstracts
Technology Research Database
Advanced Technologies Database with Aerospace
DatabaseTitle CrossRef
Technology Research Database
Advanced Technologies Database with Aerospace
Electronics & Communications Abstracts
DatabaseTitleList Technology Research Database

Database_xml – sequence: 1
  dbid: RIE
  name: IEEE/IET Electronic Library (IEL) (UW System Shared)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISSN 1557-9646
EndPage 7
ExternalDocumentID 10_1109_TED_2024_3360015
10423426
Genre orig-research
GrantInformation_xml – fundername: European Union’s Horizon 2020 Research and Innovation Program through JU and France, Belgium, Czech Republic, Germany, Italy, Sweden, Switzerland, and Turkey
– fundername: ECSEL Joint Undertaking (JU)
  grantid: 101007321
GroupedDBID -~X
.DC
0R~
29I
4.4
5GY
6IK
97E
AAJGR
AARMG
AASAJ
AAWTH
ABAZT
ABQJQ
ABVLG
ACGFO
ACGFS
ACIWK
ACNCT
AENEX
AGQYO
AHBIQ
AKJIK
AKQYR
ALMA_UNASSIGNED_HOLDINGS
ASUFR
ATWAV
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
EBS
F5P
HZ~
IFIPE
IPLJI
JAVBF
LAI
M43
MS~
O9-
OCL
P2P
RIA
RIE
RNS
TAE
TN5
3EH
5VS
AAYXX
ACKIV
AETIX
AGSQL
AI.
AIBXA
ALLEH
CITATION
EJD
H~9
IAAWW
IBMZZ
ICLAB
IDIHD
IFJZH
VH1
VJK
VOH
7SP
8FD
L7M
ID FETCH-LOGICAL-c334t-446382d2f473b5090d4e6dd72b7e2e56ebd3c9d5a3914ef0b59bb0a7ce24bb393
IEDL.DBID RIE
ISICitedReferencesCount 4
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001174124400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0018-9383
IngestDate Mon Jun 30 10:17:35 EDT 2025
Sat Nov 29 04:41:48 EST 2025
Tue Nov 18 22:24:32 EST 2025
Wed Aug 27 02:17:11 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 3
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
https://doi.org/10.15223/policy-029
https://doi.org/10.15223/policy-037
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c334t-446382d2f473b5090d4e6dd72b7e2e56ebd3c9d5a3914ef0b59bb0a7ce24bb393
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ORCID 0000-0001-9362-2113
0000-0002-0083-5804
0009-0004-9023-6101
0000-0001-7602-4058
0000-0002-1853-1614
OpenAccessLink https://hdl.handle.net/11311/1263234
PQID 2933609809
PQPubID 85466
PageCount 7
ParticipantIDs crossref_primary_10_1109_TED_2024_3360015
ieee_primary_10423426
proquest_journals_2933609809
crossref_citationtrail_10_1109_TED_2024_3360015
PublicationCentury 2000
PublicationDate 2024-03-01
PublicationDateYYYYMMDD 2024-03-01
PublicationDate_xml – month: 03
  year: 2024
  text: 2024-03-01
  day: 01
PublicationDecade 2020
PublicationPlace New York
PublicationPlace_xml – name: New York
PublicationTitle IEEE transactions on electron devices
PublicationTitleAbbrev TED
PublicationYear 2024
Publisher IEEE
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml – name: IEEE
– name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
References ref13
ref15
Li (ref6) 2018; 1
ref20
ref11
ref22
ref10
ref21
Song (ref12) 2023; 44
ref2
ref1
ref17
ref16
Cao (ref14) 2022; 12
ref19
ref18
ref7
ref9
ref4
ref3
Mahmoodi (ref5)
Kingma (ref23) 2014
Pérez (ref8) 2019; 214
References_xml – volume: 214
  start-page: 104
  year: 2019
  ident: ref8
  article-title: Analysis of the statistics of device-to-device and cycle-to-cycle variability in TiN/Ti/Al:HfO2/TiN RRAMs
  publication-title: Microelectronic Eng.
  doi: 10.1016/j.mee.2019.05.004
– volume: 44
  start-page: 1280
  issue: 8
  year: 2023
  ident: ref12
  article-title: Mitigate IR-drop effect by modulating neuron activation functions for implementing neural networks on memristor crossbar arrays
  publication-title: IEEE Electron Device Lett.
  doi: 10.1109/LED.2023.3285916
– ident: ref16
  doi: 10.1109/5.726791
– ident: ref22
  doi: 10.1109/TED.2021.3095433
– ident: ref7
  doi: 10.1109/IEDM.2007.4419107
– ident: ref17
  doi: 10.1109/ESSCIRC.2007.4430310
– ident: ref4
  doi: 10.1109/iedm19573.2019.8993599
– ident: ref21
  doi: 10.3389/fnins.2020.00634
– ident: ref1
  doi: 10.1038/s41928-018-0092-2
– ident: ref2
  doi: 10.3389/fnins.2016.00333
– ident: ref18
  doi: 10.1109/TED.2021.3089995
– volume: 1
  start-page: 52
  issue: 1
  year: 2018
  ident: ref6
  article-title: Analogue signal and image processing with large memristor crossbars
  publication-title: Nature Electron.
  doi: 10.1038/s41928-017-0002-z
– ident: ref10
  doi: 10.1116/1.1642639
– ident: ref9
  doi: 10.1109/ted.2022.3169112
– start-page: 14
  volume-title: IEDM Tech. Dig.
  ident: ref5
  article-title: An analog neuro-optimizer with adaptable annealing based on 64 × 64 0T1R crossbar circuit
– ident: ref11
  doi: 10.1109/TED.2022.3141987
– ident: ref15
  doi: 10.1109/IRPS48227.2022.9764486
– ident: ref19
  doi: 10.1145/3316781.3317872
– ident: ref3
  doi: 10.1109/ISSCC19947.2020.9062953
– ident: ref13
  doi: 10.1002/advs.202105784
– year: 2014
  ident: ref23
  article-title: Adam: A method for stochastic optimization
  publication-title: arXiv:1412.6980
– volume: 12
  start-page: 436
  issue: 2
  year: 2022
  ident: ref14
  article-title: Parasitic-aware modeling and neural network training scheme for energy-efficient processing-in-memory with resistive crossbar array
  publication-title: IEEE J. Emerg. Sel. Topics Circuits Syst.
  doi: 10.1109/JETCAS.2022.3172170
– ident: ref20
  doi: 10.1109/TED.2018.2865352
SSID ssj0016442
Score 2.4722626
Snippet In-memory computing (IMC) can accelerate data-intensive tasks, such as matrix-vector multiplication (MVM) or artificial neural networks (ANNs) inference, by...
SourceID proquest
crossref
ieee
SourceType Aggregation Database
Enrichment Source
Index Database
Publisher
StartPage 1
SubjectTerms Accuracy
Arrays
Artificial neural networks
Deep learning
emerging memory technologies
Energy consumption
hardware accelerator
in-memory computing (IMC)
Logic gates
Mathematical analysis
Matrix algebra
Microprocessors
Programming
Resistance
resistive switching memory (RRAM)
Transistors
Voltage
Wires
Title Compact Modeling and Mitigation of Parasitics in Crosspoint Accelerators of Neural Networks
URI https://ieeexplore.ieee.org/document/10423426
https://www.proquest.com/docview/2933609809
Volume 71
WOSCitedRecordID wos001174124400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE/IET Electronic Library (IEL) (UW System Shared)
  customDbUrl:
  eissn: 1557-9646
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0016442
  issn: 0018-9383
  databaseCode: RIE
  dateStart: 19630101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LS8NAEF5s8aAHnxWrVfbgxUPaNJtks8dSLB609KBS8BD2MYGCJKVJ_f3ObtJSEQVvOcwmy05mvpmdFyF3Eq02rbLYQ6hLvFBGgScAlaFQiIUgeajc1cDbE59Ok_lczJpidVcLAwAu-Qz69tHF8k2h1_aqDCUcwR8hpUVanMd1sdY2ZIDAXrcGH6IEo9-1iUn6YoA6AD3BIOwzZvE9-oZBbqjKD03s4GVy_M-NnZCjxo6ko5rxp2QP8jNyuNNd8Jy8O1nXFbXzzmzVOZW5oc-LuqtGkdMiozO5kjZrS5d0kdOx3e6yWOQVHWmNgORi8KUltE088HvTOmu87JDXycPL-NFrZil4mrGw8tDrY0lggizkTKGR4JsQYmN4oDgEEMWgDNPCRJKJYQiZryKhlC-5hiBUigl2Qdp5kcMlodmQgeaxDjiaIpHxlRJDIQB1B9IbI7tksDndVDeNxu28i4_UORy-SJEfqeVH2vCjS-63K5Z1k40_aDv2_Hfo6qPvkt6Gg2kjhmWKtgyuEokvrn5Zdk0O7NvrrLIeaVerNdyQff1ZLcrVrfvDvgD0X81w
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LS8QwEB58gXrwueL6zMGLh2o3Sbebo4iiuC4eVAQPJY8pLEgr2-rvd5J2RREFbz1MSMh05pvJvACONFlt1uT9iKBuEEmd8EghKUNlCAtRp9KEp4HHYToaDZ6e1F1brB5qYRAxJJ_hif8MsXxX2jf_VEYSTuBPkDIL84mUPG7KtT6DBgTtTXPwHskweV7TqGSsTkkLkC_I5YkQHuGTbygUxqr80MUBYC5X_3m0NVhpLUl21rB-HWaw2IDlL_0FN-E5SLutmZ945uvOmS4cux03fTXKgpU5u9MT7fO2bMXGBTv3x30tx0XNzqwlSApR-MoT-jYetN-oyRuvOvBweXF_fhW10xQiK4SsI_L7xIA7nstUGDITYiex71zKTYockz4aJ6xyiRaqJzGPTaKMiXVqkUtjhBJbMFeUBW4Dy3sCbdq3PCVjJHGxMaqnFJL2IHrndBdOp7eb2bbVuJ948ZIFlyNWGfEj8_zIWn504fhzxWvTZuMP2o6__y90zdV3YW_KwawVxCoja4ZWqUGsdn5ZdgiLV_e3w2x4PbrZhSW_U5Njtgdz9eQN92HBvtfjanIQ_rYPzebQtw
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Compact+Modeling+and+Mitigation+of+Parasitics+in+Crosspoint+Accelerators+of+Neural+Networks&rft.jtitle=IEEE+transactions+on+electron+devices&rft.au=Lepri%2C+N.&rft.au=Glukhov%2C+A.&rft.au=Mannocci%2C+P.&rft.au=Porzani%2C+M.&rft.date=2024-03-01&rft.pub=IEEE&rft.issn=0018-9383&rft.spage=1&rft.epage=7&rft_id=info:doi/10.1109%2FTED.2024.3360015&rft.externalDocID=10423426
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0018-9383&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0018-9383&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0018-9383&client=summon