COSIME: FeFET based Associative Memory for In-Memory Cosine Similarity Search

In a number of machine learning models, an input query is searched across the trained class vectors to find the closest feature class vector in cosine similarity metric. However, performing the cosine similarities between the vectors in Von-Neumann machines involves a large number of multiplications...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2022 IEEE/ACM International Conference On Computer Aided Design (ICCAD) s. 1 - 9
Hlavní autoři: Liu, Che-Kai, Chen, Haobang, Imani, Mohsen, Ni, Kai, Kazemi, Arman, Laguna, Ann Franchesca, Niemier, Michael, Hu, Xiaobo Sharon, Zhao, Liang, Zhuo, Cheng, Yin, Xunzhao
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: ACM 29.10.2022
Témata:
ISSN:1558-2434
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract In a number of machine learning models, an input query is searched across the trained class vectors to find the closest feature class vector in cosine similarity metric. However, performing the cosine similarities between the vectors in Von-Neumann machines involves a large number of multiplications, Euclidean normalizations and division operations, thus incurring heavy hardware energy and latency overheads. Moreover, due to the memory wall problem that presents in the conventional architecture, frequent cosine similarity-based searches (CSSs) over the class vectors requires a lot of data movements, limiting the throughput and efficiency of the system. To overcome the aforementioned challenges, this paper introduces COSIME, a general in-memory associative memory (AM) engine based on the ferroelectric FET (FeFET) device for efficient CSS. By leveraging the one-transistor AND gate function of FeFET devices, current-based translinear analog circuit and winner-take-all (WTA) circuitry, COSIME can realize parallel in-memory CSS across all the entries in a memory block, and output the closest word to the input query in cosine similarity metric. Evaluation results at the array level suggest that the proposed COSIME design achieves 333× and 90.5× latency and energy improvements, respectively, and realizes better classification accuracy when compared with an AM design implementing approximated CSS. The proposed in-memory computing fabric is evaluated for an HDC problem, showcasing that COSIME can achieve on average 47.1× and 98.5× speedup and energy efficiency improvements compared with an GPU implementation.
AbstractList In a number of machine learning models, an input query is searched across the trained class vectors to find the closest feature class vector in cosine similarity metric. However, performing the cosine similarities between the vectors in Von-Neumann machines involves a large number of multiplications, Euclidean normalizations and division operations, thus incurring heavy hardware energy and latency overheads. Moreover, due to the memory wall problem that presents in the conventional architecture, frequent cosine similarity-based searches (CSSs) over the class vectors requires a lot of data movements, limiting the throughput and efficiency of the system. To overcome the aforementioned challenges, this paper introduces COSIME, a general in-memory associative memory (AM) engine based on the ferroelectric FET (FeFET) device for efficient CSS. By leveraging the one-transistor AND gate function of FeFET devices, current-based translinear analog circuit and winner-take-all (WTA) circuitry, COSIME can realize parallel in-memory CSS across all the entries in a memory block, and output the closest word to the input query in cosine similarity metric. Evaluation results at the array level suggest that the proposed COSIME design achieves 333× and 90.5× latency and energy improvements, respectively, and realizes better classification accuracy when compared with an AM design implementing approximated CSS. The proposed in-memory computing fabric is evaluated for an HDC problem, showcasing that COSIME can achieve on average 47.1× and 98.5× speedup and energy efficiency improvements compared with an GPU implementation.
Author Yin, Xunzhao
Chen, Haobang
Hu, Xiaobo Sharon
Kazemi, Arman
Imani, Mohsen
Zhao, Liang
Liu, Che-Kai
Zhuo, Cheng
Ni, Kai
Laguna, Ann Franchesca
Niemier, Michael
Author_xml – sequence: 1
  givenname: Che-Kai
  surname: Liu
  fullname: Liu, Che-Kai
  organization: Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China
– sequence: 2
  givenname: Haobang
  surname: Chen
  fullname: Chen, Haobang
  organization: Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China
– sequence: 3
  givenname: Mohsen
  surname: Imani
  fullname: Imani, Mohsen
  organization: University of California,Department of Computer Science,Irvine,CA,USA
– sequence: 4
  givenname: Kai
  surname: Ni
  fullname: Ni, Kai
  organization: Rochester Institute of Technology,Department of Electrical and Microelectronic Engineering,NY,USA
– sequence: 5
  givenname: Arman
  surname: Kazemi
  fullname: Kazemi, Arman
  organization: University of Notre Dame,Department of Computer Science and Engineering,IN,USA
– sequence: 6
  givenname: Ann Franchesca
  surname: Laguna
  fullname: Laguna, Ann Franchesca
  organization: De La Salle University,Department of Computer Technology,Manilla,Philippines
– sequence: 7
  givenname: Michael
  surname: Niemier
  fullname: Niemier, Michael
  organization: University of Notre Dame,Department of Computer Science and Engineering,IN,USA
– sequence: 8
  givenname: Xiaobo Sharon
  surname: Hu
  fullname: Hu, Xiaobo Sharon
  organization: University of Notre Dame,Department of Computer Science and Engineering,IN,USA
– sequence: 9
  givenname: Liang
  surname: Zhao
  fullname: Zhao, Liang
  organization: Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China
– sequence: 10
  givenname: Cheng
  surname: Zhuo
  fullname: Zhuo, Cheng
  email: czhuo@zju.edu.cn
  organization: Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China
– sequence: 11
  givenname: Xunzhao
  surname: Yin
  fullname: Yin, Xunzhao
  email: xzyin1@zju.edu.cn
  organization: Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China
BookMark eNotjM1qAjEURtPSQq113U0XeYGxuUlufrqTQdsBxYXtWuLkhqboTJlIwbevoKvzHT44j-yu6zti7BnEFEDjq0LhFMqpQu01yBs28dadD6G8BKtv2QgQXSW10g9sUsqPEEI6C9aKEVvV602zmr_xBS3mn3wXCkU-K6VvczjmP-IrOvTDiad-4E1XXa3uS-6Ib_Ih78OQjye-oTC030_sPoV9ocmVY_Z1rtYf1XL93tSzZRWkdscqxaRSdF4aZUPyO9TOGH-eAk10GiCYpOXOYWtamxQCtVF6RBFjQkJQY_Zy6WYi2v4O-RCG0xaEMF5rUP8--k5-
ContentType Conference Proceeding
DBID 6IE
6IH
CBEJK
RIE
RIO
DOI 10.1145/3508352.3549412
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan (POP) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP) 1998-present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISBN 9781450392174
1450392172
EISSN 1558-2434
EndPage 9
ExternalDocumentID 10069441
Genre orig-research
GrantInformation_xml – fundername: Semiconductor Research Corporation
  funderid: 10.13039/100000028
– fundername: National Science Foundation
  funderid: 10.13039/100000001
GroupedDBID 6IE
6IF
6IH
6IL
6IN
AAWTH
ABLEC
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
FEDTE
IEGSK
IJVOP
M43
OCL
RIE
RIL
RIO
ID FETCH-LOGICAL-a248t-fdf3fd892637af9b5486697af056d8411a6f42b85c6c7f351ecd29550ddf5e513
IEDL.DBID RIE
ISICitedReferencesCount 22
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000981574300136&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
IngestDate Wed Aug 27 02:46:16 EDT 2025
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-a248t-fdf3fd892637af9b5486697af056d8411a6f42b85c6c7f351ecd29550ddf5e513
PageCount 9
ParticipantIDs ieee_primary_10069441
PublicationCentury 2000
PublicationDate 2022-Oct.-29
PublicationDateYYYYMMDD 2022-10-29
PublicationDate_xml – month: 10
  year: 2022
  text: 2022-Oct.-29
  day: 29
PublicationDecade 2020
PublicationTitle 2022 IEEE/ACM International Conference On Computer Aided Design (ICCAD)
PublicationTitleAbbrev ICCAD
PublicationYear 2022
Publisher ACM
Publisher_xml – name: ACM
SSID ssj0002871770
ssj0020286
Score 2.3330417
Snippet In a number of machine learning models, an input query is searched across the trained class vectors to find the closest feature class vector in cosine...
SourceID ieee
SourceType Publisher
StartPage 1
SubjectTerms Associative memory
Energy efficiency
Graphics processing units
Measurement
Nonvolatile memory
Support vector machines
Throughput
Title COSIME: FeFET based Associative Memory for In-Memory Cosine Similarity Search
URI https://ieeexplore.ieee.org/document/10069441
WOSCitedRecordID wos000981574300136&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
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07T8MwELagYoCFVxFveWB1IbbjB2vVikq0VCqgblVin6UOpKi0lfj3-NxQWBjYHA9WdLZz33e5746QGwHBWQWceRAFi_iWs5I7w7QJBl2iUyne8fqoBwMzHtthLVZPWhgASMln0MJh-pfvZ26JobJ4w1GmiTL1ba31Wqy1Cagg9Nd4-Gq2FSdUXcsnk_mtyBPYaIlIiCT2n_zVTCX5ku7-P9_igDR_VHl0uPE3h2QLqiOy96ug4DHpt59GvX7nnnah23mm6KI83ezACmgfE2s_aUSqtFex-qk9w-R3Opq-TSPPjbCcrpOQm-QlrtJ-YHXDBFZwaRYs-CCCN5YroYtgy8hGlLJxGFGONzLLChUkL03ulNNB5Bk4z23kKN6HHPJMnJBGNavglFDLC2QyBgGHdHfCqDLyaFlgow4TApyRJlpm8r6uiTH5Nsr5H_MXZJejcCB-9bm9JI3FfAlXZMetFtOP-XXayS8OOptr
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV09T8MwELUQIAELX0V844HVhdiOY7NWrVrRlEotqFuV2GepAykqbSX-Pb40lC4MbLaHKLITvffO9-4IuRfgrVHAmQORscBvOcu51SzRXiMkWlXGO966Sa-nRyPTr8zqpRcGAMrkM6jjsLzLd1O7wFBZ-MPRpok29Z1YSh6t7FrrkAqS_wQ_v0pvhQVVVfOJZPwg4pJu1EWQRBI7UG60UynRpHX4z_c4IrVfXx7trxHnmGxBcUIONkoKnpK08TLopM0n2oJWc0gRpBxdn8ESaIqptV80cFXaKVg1a0wx_Z0OJu-ToHQDMaerNOQaeQ1PabRZ1TKBZVzqOfPOC--04UokmTd50CNKmTAMPMdpGUWZ8pLnOrbKJl7EEVjHTVApzvkY4kicke1iWsA5oYZnqGU0Ug5pH4VWeVDSMsNWHdp7uCA13Jnxx6oqxvhnUy7_WL8je-1h2h13O73nK7LP0UYQMICba7I9ny3ghuza5XzyObstT_UbWCCesg
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%3Abook&rft.genre=proceeding&rft.title=2022+IEEE%2FACM+International+Conference+On+Computer+Aided+Design+%28ICCAD%29&rft.atitle=COSIME%3A+FeFET+based+Associative+Memory+for+In-Memory+Cosine+Similarity+Search&rft.au=Liu%2C+Che-Kai&rft.au=Chen%2C+Haobang&rft.au=Imani%2C+Mohsen&rft.au=Ni%2C+Kai&rft.date=2022-10-29&rft.pub=ACM&rft.eissn=1558-2434&rft.spage=1&rft.epage=9&rft_id=info:doi/10.1145%2F3508352.3549412&rft.externalDocID=10069441