Performance Characterization of Popular DNN Models on Out-of-Order CPUs

DNN popularity, which is driving advances in a growing number of fields, has increased the amount of computing resources running this kind of applications at an unprecedent rate. Specialized hardware, such as GPUs or ASIC-based accelerators, has been the preferred platform to run these applications....

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2023 32nd International Conference on Parallel Architectures and Compilation Techniques (PACT) s. 199 - 210
Hlavní autoři: Prieto, Pablo, Abad, Pablo, Gregorio, Jose Angel, Puente, Valentin
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 21.10.2023
Témata:
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract DNN popularity, which is driving advances in a growing number of fields, has increased the amount of computing resources running this kind of applications at an unprecedent rate. Specialized hardware, such as GPUs or ASIC-based accelerators, has been the preferred platform to run these applications. However, the ubiquity of DNN models is rapidly extending the presence of this software to general-purpose CPUs. For this reason, there is a pressing need to gain understanding of the main features of state-of-the-art DNN models to adapt CPU microarchitecture accordingly. In this paper we investigated a representative set of DNN models and, based on data collected from real hardware, we evaluated how efficiently they utilize the underlying system. We analyzed overall system performance, as well as the amount of vectorization provided by CPU-optimized frameworks. We quantified the performance loss caused by processor backend, and the contribution of memory hierarchy and functional units to it. We compared the backend utilization of DNN applications to popular benchmarks such as SPEC CPU2017 and found a lower balance in the use of the elements that make up the processor microarchitecture. Although many workloads seem to be constrained by functional unit availability, in a significant group of applications we found a non-negligible impact of memory hierarchy on performance.
AbstractList DNN popularity, which is driving advances in a growing number of fields, has increased the amount of computing resources running this kind of applications at an unprecedent rate. Specialized hardware, such as GPUs or ASIC-based accelerators, has been the preferred platform to run these applications. However, the ubiquity of DNN models is rapidly extending the presence of this software to general-purpose CPUs. For this reason, there is a pressing need to gain understanding of the main features of state-of-the-art DNN models to adapt CPU microarchitecture accordingly. In this paper we investigated a representative set of DNN models and, based on data collected from real hardware, we evaluated how efficiently they utilize the underlying system. We analyzed overall system performance, as well as the amount of vectorization provided by CPU-optimized frameworks. We quantified the performance loss caused by processor backend, and the contribution of memory hierarchy and functional units to it. We compared the backend utilization of DNN applications to popular benchmarks such as SPEC CPU2017 and found a lower balance in the use of the elements that make up the processor microarchitecture. Although many workloads seem to be constrained by functional unit availability, in a significant group of applications we found a non-negligible impact of memory hierarchy on performance.
Author Abad, Pablo
Puente, Valentin
Gregorio, Jose Angel
Prieto, Pablo
Author_xml – sequence: 1
  givenname: Pablo
  surname: Prieto
  fullname: Prieto, Pablo
  email: prietop@unican.es
  organization: Universidad de Cantabria,Computer Engineering Group,Santander,Spain
– sequence: 2
  givenname: Pablo
  surname: Abad
  fullname: Abad, Pablo
  email: abadp@unican.es
  organization: Universidad de Cantabria,Computer Engineering Group,Santander,Spain
– sequence: 3
  givenname: Jose Angel
  surname: Gregorio
  fullname: Gregorio, Jose Angel
  email: monaster@unican.es
  organization: Universidad de Cantabria,Computer Engineering Group,Santander,Spain
– sequence: 4
  givenname: Valentin
  surname: Puente
  fullname: Puente, Valentin
  email: vpuente@unican.es
  organization: Universidad de Cantabria,Computer Engineering Group,Santander,Spain
BookMark eNotj0FLwzAYQCMoqLP_YIf8gc7k-9ImPY6qU5hrD9t5pMkXLHTNSLuD_noHenqHBw_eI7sd40iMLaVYSSmq53Zd7wsjpV6BAFwJIaC4YVmlK4OFQAWFwnuWTVPfiUJr1FDhA9u0lEJMJzs64vWXTdbNlPofO_dx5DHwNp4vg038Zbfjn9HTMPGraC5zHkPeJE-J1-1hemJ3wQ4TZf9csMPb675-z7fN5qNeb3MLppxzZUynqbOOgjKO0JnSIwivg3SyUBpA6MrJkkhCWRFB5wGCJUXolUPABVv-dXsiOp5Tf7Lp-ygFluo6j791_00T
CODEN IEEPAD
ContentType Conference Proceeding
DBID 6IE
6IL
CBEJK
RIE
RIL
DOI 10.1109/PACT58117.2023.00025
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
EISBN 9798350342543
EndPage 210
ExternalDocumentID 10364581
Genre orig-research
GrantInformation_xml – fundername: Spanish Government (Agencia Estatal de Investigación)
  grantid: PID2019-110051GB-100,PID2022-139664NB-100
  funderid: 10.13039/501100011033
GroupedDBID 6IE
6IL
ACM
ALMA_UNASSIGNED_HOLDINGS
APO
CBEJK
LHSKQ
RIE
RIL
ID FETCH-LOGICAL-a286t-488b7ebacef48ce3c86d320d7f1c154722079c16ee1269ee2bd22fae4e3d4c323
IEDL.DBID RIE
ISICitedReferencesCount 1
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001165646800017&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 02:24:17 EDT 2025
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-a286t-488b7ebacef48ce3c86d320d7f1c154722079c16ee1269ee2bd22fae4e3d4c323
PageCount 12
ParticipantIDs ieee_primary_10364581
PublicationCentury 2000
PublicationDate 2023-Oct.-21
PublicationDateYYYYMMDD 2023-10-21
PublicationDate_xml – month: 10
  year: 2023
  text: 2023-Oct.-21
  day: 21
PublicationDecade 2020
PublicationTitle 2023 32nd International Conference on Parallel Architectures and Compilation Techniques (PACT)
PublicationTitleAbbrev PACT
PublicationYear 2023
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssib057737293
Score 2.2425725
Snippet DNN popularity, which is driving advances in a growing number of fields, has increased the amount of computing resources running this kind of applications at...
SourceID ieee
SourceType Publisher
StartPage 199
SubjectTerms Adaptation models
Analytical models
Artificial neural networks
Deep Neural Networks
General Purpose CPU
Gold
Metrics
Microarchitecture
Silver
System performance
Title Performance Characterization of Popular DNN Models on Out-of-Order CPUs
URI https://ieeexplore.ieee.org/document/10364581
WOSCitedRecordID wos001165646800017&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/eLvHCXMwlV07T8MwELagYmACRBFveWB1SewkjkcUKExphlbqVvlxlpBQg9qU31-f-2JhYLPswdLZd2f57vs-Qp6kkCa24hifSRaiJGemNBnLMov5SeVeuCg2Ieu6nE5VswWrRywMAMTmMxjgMNbyXWtX-FUWPByLZgi0Ppay2IC1dpcnlyi4osQWHpcm6rl5qcY5AikHqBGOTIUoiP1LRCXmkOHZP3c_J_0DGo82-zxzQY5gfknem0PHP632rMsbUCVtPW2iMteCvtY1RcGzryUNC6NVx1rPRki4SatmsuyTyfBtXH2wrSgC07wsOhYczkgw2oLPSgvCloUTPHHSpzY8hyTniVQ2LQBSXigAbhznXkMGwmVWcHFFevN2DteEYvnWFco4n_twNInWXpcuD-lLlUYX7ob00Qqz7w3vxWxngNs_5u_IKRoaIztP70mvW6zggZzYn-5zuXiMp7UGl5KVWw
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlZ09T8MwEIYtVJBgAkQR33hgdUnsJI5HFChFlDRDK3WrYvssIaEGtSm_n5ybUhYGNsseLJ0_zvLdew8hd1JI7VNxtIska25JznSqIxZFBv2Tip2wHjYh8zydTlXRitW9FgYAfPIZ9LDpY_m2Miv8KmtOOAbNUGi9i-isVq612T6xROSKEq1ALgzUffGQjWOUUvaQEo61ChGJ_Quj4r1I__Cf8x-R7laPR4sfT3NMdmB-Qp6Lbc4_zX7qLq9llbRytPBsrgV9zHOKyLOPJW0GRquaVY6NsOQmzYrJsksm_adxNmAtFoGVPE1q1hw5LUGXBlyUGhAmTazggZUuNM2DSHIeSGXCBCDkiQLg2nLuSohA2MgILk5JZ17N4YxQDODaRGnrYtcsTlCWrkxt3Dgwleoyseeki1aYfa4rX8w2Brj4o_-W7A_Gb8PZ8CV_vSQHaHS853l4RTr1YgXXZM981e_LxY1fuW-RYZik
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=2023+32nd+International+Conference+on+Parallel+Architectures+and+Compilation+Techniques+%28PACT%29&rft.atitle=Performance+Characterization+of+Popular+DNN+Models+on+Out-of-Order+CPUs&rft.au=Prieto%2C+Pablo&rft.au=Abad%2C+Pablo&rft.au=Gregorio%2C+Jose+Angel&rft.au=Puente%2C+Valentin&rft.date=2023-10-21&rft.pub=IEEE&rft.spage=199&rft.epage=210&rft_id=info:doi/10.1109%2FPACT58117.2023.00025&rft.externalDocID=10364581