AGNI: In-Situ, Iso-Latency Stochastic-to-Binary Number Conversion for In-DRAM Deep Learning

Recent years have seen a rapid increase in research activity in the field of DRAM-based Processing-In-Memory (PIM) accelerators, where the analog computing capability of DRAM is employed by minimally changing the inherent structure of DRAM peripherals to accelerate various data-centric applications....

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Proceedings / IEEE International Symposium on Quality Electronic Design s. 1 - 8
Hlavní autoři: Shivanandamurthy, Supreeth Mysore, Vatsavai, Sairam Sri, Thakkar, Ishan, Salehi, Sayed Ahmad
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 05.04.2023
Témata:
ISSN:1948-3295
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Recent years have seen a rapid increase in research activity in the field of DRAM-based Processing-In-Memory (PIM) accelerators, where the analog computing capability of DRAM is employed by minimally changing the inherent structure of DRAM peripherals to accelerate various data-centric applications. Several DRAM-based PIM accelerators for Convolutional Neural Networks (CNNs) have also been reported. Among these, the accelerators leveraging in-DRAM stochastic arithmetic have shown manifold improvements in processing latency and throughput, due to the ability of stochastic arithmetic to convert multiplications into simple bit-wise logical AND operations. However, the use of in-DRAM stochastic arithmetic for CNN acceleration requires frequent stochastic to binary number conversions. For that, prior works employ full adder-based or serial counter-based in-DRAM circuits. These circuits consume large area and incur long latency. Their in-DRAM implementations also require heavy modifications in DRAM peripherals, which significantly diminishes the benefits of using stochastic arithmetic in these accelerators. To address these shortcomings, this paper presents a new substrate for in-DRAM stochastic-to-binary number conversion called AGNI. AGNI makes minor modifications in DRAM peripherals using pass transistors, capacitors, encoders, and charge pumps, and re-purposes the sense amplifiers as voltage comparators, to enable in-situ binary conversion of input statistic operands of different sizes with iso latency. Our evaluations, based on detailed SPICE simulations (https://github.com/uky-UCAT/AGNI_SPICE.git), show that AGNI can achieve savings of at least 8× in area, at least 28× energy-delay product (EDP), and at least 21 in area × latency, compared to two in-DRAM stochastic-to-binary conversion circuits from prior works. These circuit-level benefits are demonstrated to propagate at the system-level to achieve at least 3.9× gain in performance across four deep CNN models.
AbstractList Recent years have seen a rapid increase in research activity in the field of DRAM-based Processing-In-Memory (PIM) accelerators, where the analog computing capability of DRAM is employed by minimally changing the inherent structure of DRAM peripherals to accelerate various data-centric applications. Several DRAM-based PIM accelerators for Convolutional Neural Networks (CNNs) have also been reported. Among these, the accelerators leveraging in-DRAM stochastic arithmetic have shown manifold improvements in processing latency and throughput, due to the ability of stochastic arithmetic to convert multiplications into simple bit-wise logical AND operations. However, the use of in-DRAM stochastic arithmetic for CNN acceleration requires frequent stochastic to binary number conversions. For that, prior works employ full adder-based or serial counter-based in-DRAM circuits. These circuits consume large area and incur long latency. Their in-DRAM implementations also require heavy modifications in DRAM peripherals, which significantly diminishes the benefits of using stochastic arithmetic in these accelerators. To address these shortcomings, this paper presents a new substrate for in-DRAM stochastic-to-binary number conversion called AGNI. AGNI makes minor modifications in DRAM peripherals using pass transistors, capacitors, encoders, and charge pumps, and re-purposes the sense amplifiers as voltage comparators, to enable in-situ binary conversion of input statistic operands of different sizes with iso latency. Our evaluations, based on detailed SPICE simulations (https://github.com/uky-UCAT/AGNI_SPICE.git), show that AGNI can achieve savings of at least 8× in area, at least 28× energy-delay product (EDP), and at least 21 in area × latency, compared to two in-DRAM stochastic-to-binary conversion circuits from prior works. These circuit-level benefits are demonstrated to propagate at the system-level to achieve at least 3.9× gain in performance across four deep CNN models.
Author Salehi, Sayed Ahmad
Vatsavai, Sairam Sri
Thakkar, Ishan
Shivanandamurthy, Supreeth Mysore
Author_xml – sequence: 1
  givenname: Supreeth Mysore
  surname: Shivanandamurthy
  fullname: Shivanandamurthy, Supreeth Mysore
  organization: University of Kentucky,Department of Electrical and Computer Engineering,Lexington,KY,USA,40506
– sequence: 2
  givenname: Sairam Sri
  surname: Vatsavai
  fullname: Vatsavai, Sairam Sri
  organization: University of Kentucky,Department of Electrical and Computer Engineering,Lexington,KY,USA,40506
– sequence: 3
  givenname: Ishan
  surname: Thakkar
  fullname: Thakkar, Ishan
  organization: University of Kentucky,Department of Electrical and Computer Engineering,Lexington,KY,USA,40506
– sequence: 4
  givenname: Sayed Ahmad
  surname: Salehi
  fullname: Salehi, Sayed Ahmad
  organization: University of Kentucky,Department of Electrical and Computer Engineering,Lexington,KY,USA,40506
BookMark eNo1kMtOwzAURA0Cibb0D1j4A3C5147jmF3pi0ihCAorFlVibsCI2lWSIvXvKQJWs5qjOdNnJyEGYowjjBDBXuWrh9lUGyvNSIJUIwSUVgEesaE1NlMalEqMVseshzbJhJJWn7F-234AJFqbrMdexotlfs3zIFa-213yvI2iKDsKbs9XXXTvZdt5J7oobnwomz1f7jYVNXwSwxc1rY-B17H56U8fx3d8SrTlBZVN8OHtnJ3W5WdLw78csOf57GlyK4r7RT4ZF8JLSDqBVSqdTWVWg7UpJaasUTtwGrQxkEpJUr0afdBBaSh1iFhlQLoyTlFV1WrALn65nojW28ZvDkPX_2eob9OZU9U
ContentType Conference Proceeding
DBID 6IE
6IL
CBEJK
RIE
RIL
DOI 10.1109/ISQED57927.2023.10129301
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP All) 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 9798350334753
EISSN 1948-3295
EndPage 8
ExternalDocumentID 10129301
Genre orig-research
GroupedDBID 6IE
6IF
6IH
6IK
6IL
6IN
AAJGR
AAWTH
ABLEC
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
IEGSK
IPLJI
OCL
RIE
RIL
ID FETCH-LOGICAL-i204t-1b62c9628f0996e47af15c0c505770622e23d75835127e6c111b80e5b7c3ebbf3
IEDL.DBID RIE
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001013619400003&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 02:58:30 EDT 2025
IsPeerReviewed false
IsScholarly false
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i204t-1b62c9628f0996e47af15c0c505770622e23d75835127e6c111b80e5b7c3ebbf3
PageCount 8
ParticipantIDs ieee_primary_10129301
PublicationCentury 2000
PublicationDate 2023-April-5
PublicationDateYYYYMMDD 2023-04-05
PublicationDate_xml – month: 04
  year: 2023
  text: 2023-April-5
  day: 05
PublicationDecade 2020
PublicationTitle Proceedings / IEEE International Symposium on Quality Electronic Design
PublicationTitleAbbrev ISQED
PublicationYear 2023
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0045578
Score 1.8424108
Snippet Recent years have seen a rapid increase in research activity in the field of DRAM-based Processing-In-Memory (PIM) accelerators, where the analog computing...
SourceID ieee
SourceType Publisher
StartPage 1
SubjectTerms Convolutional neural networks
processing-in-memory
Random access memory
SPICE
Stochastic processes
stochastic to binary conversion
Throughput
Transistors
Voltage
Title AGNI: In-Situ, Iso-Latency Stochastic-to-Binary Number Conversion for In-DRAM Deep Learning
URI https://ieeexplore.ieee.org/document/10129301
WOSCitedRecordID wos001013619400003&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/eLvHCXMwlV05T8MwGLWgYoCFq4hbHhhxm9hxHLOVHhCpRIUCqsRQxUegS1K1KRL_HtttOAYGNsuSZemz9R32994D4MJXRFGlFaKCpijIcIh4EAmUMW5KZo8Kz2ksPfdZkkSjER-swOoOC6O1ds1numGH7i9fFXJhn8qalouKE4vWWmcsXIK1KrcbUHP3qlYdjzfj4X23QxnHrGEVwhvV2l8qKi6I9Lb_uf0OqH_D8eDgK9DsgjWd74GtH0yC--CldZPEVzDO0XBSLi5hPC9QP7X58AccloV8Sy0fMyoLdO0AuDBxSiCwbZvO3YsZNNmrXd95aN3BjtZTuKJefa2Dp173sX2LVroJaIK9oES-CLHkIY4yk_6FOmBp5lPpSVuLMC_EWGOiTJ1ATLBnOpTG3YnI01QwSbQQGTkAtbzI9SGANv8SJJBSOQV1GUkuiEjTKFC-8U_ZEahbO42nS2qMcWWi4z_mT8CmPQ3X-kJPQa2cLfQZ2JDv5WQ-O3cH-gktcp6s
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV05T8MwGLUQIAELVxE3HhhxmzhxErOVHjQijQotqBJDFR-BLknVpkj8e2yTcAwMbJYlS9Zn6zvs770HwKUtHEGEFIgwkiA3xR6ibsBQ6lNVMluEWUZj6Sny4zgYj-mgBKsbLIyU0jSfyboemr98kfOlfipraC4q6mi01pqWzirhWpXjdYm6fVWzjkUb4fC-0yY-xX5da4TXq9W_dFRMGOlu_3MDO6D2DciDg69QswtWZLYHtn5wCe6D5-ZtHF7DMEPDabG8guEiR1GiM-J3OCxy_ppoRmZU5OjGQHBhbLRAYEu3nZs3M6jyV72-_dDsw7aUM1iSr77UwGO3M2r1UKmcgKbYcgtkMw9z6uEgVQmgJ10_SW3CLa6rEd_yMJbYEapScFS496XHlcNjgSUJ87kjGUudA7Ca5Zk8BFBnYMxxORdGQ50HnDKHJUngClt5qPQI1LSdJrNPcoxJZaLjP-YvwEZv1I8mURjfnYBNfTKmEYacgtVivpRnYJ2_FdPF_Nwc7gekvqH1
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=Proceedings+%2F+IEEE+International+Symposium+on+Quality+Electronic+Design&rft.atitle=AGNI%3A+In-Situ%2C+Iso-Latency+Stochastic-to-Binary+Number+Conversion+for+In-DRAM+Deep+Learning&rft.au=Shivanandamurthy%2C+Supreeth+Mysore&rft.au=Vatsavai%2C+Sairam+Sri&rft.au=Thakkar%2C+Ishan&rft.au=Salehi%2C+Sayed+Ahmad&rft.date=2023-04-05&rft.pub=IEEE&rft.eissn=1948-3295&rft.spage=1&rft.epage=8&rft_id=info:doi/10.1109%2FISQED57927.2023.10129301&rft.externalDocID=10129301