{"id":"https://openalex.org/W2736230459","doi":"https://doi.org/10.1109/ispass.2017.7975270","title":"Performance analysis of CNN frameworks for GPUs","display_name":"Performance analysis of CNN frameworks for GPUs","publication_year":2017,"publication_date":"2017-04-01","ids":{"openalex":"https://openalex.org/W2736230459","doi":"https://doi.org/10.1109/ispass.2017.7975270","mag":"2736230459"},"language":"en","primary_location":{"id":"doi:10.1109/ispass.2017.7975270","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ispass.2017.7975270","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2017 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5057343217","display_name":"Heehoon Kim","orcid":"https://orcid.org/0000-0002-5392-4413"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Heehoon Kim","raw_affiliation_strings":["Department of Computer Science and Engineering"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025937126","display_name":"Hyoungwook Nam","orcid":null},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Hyoungwook Nam","raw_affiliation_strings":["College of Liberal Studies, Seoul National University, Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"College of Liberal Studies, Seoul National University, Korea","institution_ids":["https://openalex.org/I139264467"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102939095","display_name":"Wookeun Jung","orcid":"https://orcid.org/0000-0001-6750-6979"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wookeun Jung","raw_affiliation_strings":["Department of Computer Science and Engineering"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100767182","display_name":"Jaejin Lee","orcid":"https://orcid.org/0000-0003-4638-8170"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jaejin Lee","raw_affiliation_strings":["Department of Computer Science and Engineering"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":4.5262,"has_fulltext":false,"cited_by_count":112,"citation_normalized_percentile":{"value":0.96843684,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"55","last_page":"64"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":1.0,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":1.0,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9995999932289124,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9994000196456909,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8897830247879028},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.7622497081756592},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.7213653326034546},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.675430417060852},{"id":"https://openalex.org/keywords/convolution","display_name":"Convolution (computer science)","score":0.633385419845581},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.6056145429611206},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5809161067008972},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5732657313346863},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.5311946272850037},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.505242645740509},{"id":"https://openalex.org/keywords/kernel","display_name":"Kernel (algebra)","score":0.49595531821250916},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.4334029257297516},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.34388160705566406},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.32902613282203674},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.31517189741134644}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8897830247879028},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.7622497081756592},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7213653326034546},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.675430417060852},{"id":"https://openalex.org/C45347329","wikidata":"https://www.wikidata.org/wiki/Q5166604","display_name":"Convolution (computer science)","level":3,"score":0.633385419845581},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.6056145429611206},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5809161067008972},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5732657313346863},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.5311946272850037},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.505242645740509},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.49595531821250916},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.4334029257297516},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.34388160705566406},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.32902613282203674},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.31517189741134644},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","level":1,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.0},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.0},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/ispass.2017.7975270","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ispass.2017.7975270","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2017 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":47,"referenced_works":["https://openalex.org/W753012316","https://openalex.org/W1487564550","https://openalex.org/W1547840952","https://openalex.org/W1568229137","https://openalex.org/W1667652561","https://openalex.org/W1686810756","https://openalex.org/W1789336918","https://openalex.org/W1872489089","https://openalex.org/W1922123711","https://openalex.org/W1995562189","https://openalex.org/W2031489346","https://openalex.org/W2097117768","https://openalex.org/W2102605133","https://openalex.org/W2112796928","https://openalex.org/W2117539524","https://openalex.org/W2119144962","https://openalex.org/W2133564696","https://openalex.org/W2145339207","https://openalex.org/W2155893237","https://openalex.org/W2163605009","https://openalex.org/W2168231600","https://openalex.org/W2172654076","https://openalex.org/W2178615544","https://openalex.org/W2181607856","https://openalex.org/W2342840547","https://openalex.org/W2384495648","https://openalex.org/W2407022425","https://openalex.org/W2949245006","https://openalex.org/W2950179405","https://openalex.org/W2952388062","https://openalex.org/W2963340555","https://openalex.org/W2963674387","https://openalex.org/W4293460823","https://openalex.org/W4302296459","https://openalex.org/W6622239757","https://openalex.org/W6632552335","https://openalex.org/W6637151318","https://openalex.org/W6637373629","https://openalex.org/W6638020065","https://openalex.org/W6639260565","https://openalex.org/W6640442106","https://openalex.org/W6674914833","https://openalex.org/W6679434410","https://openalex.org/W6684191040","https://openalex.org/W6684859321","https://openalex.org/W6685214006","https://openalex.org/W6685730785"],"related_works":["https://openalex.org/W2058965144","https://openalex.org/W2164382479","https://openalex.org/W2146343568","https://openalex.org/W98480971","https://openalex.org/W2150291671","https://openalex.org/W2013643406","https://openalex.org/W2983282793","https://openalex.org/W1973046741","https://openalex.org/W1966837078","https://openalex.org/W3157543420"],"abstract_inverted_index":{"Thanks":[0],"to":[1,61,101,109,163],"modern":[2],"deep":[3,33,138],"learning":[4,34,139],"frameworks":[5,140],"that":[6,151],"exploit":[7],"GPUs,":[8],"convolutional":[9],"neural":[10],"networks":[11],"(CNNs)":[12],"have":[13],"been":[14],"greatly":[15],"successful":[16],"in":[17,42,119,141],"visual":[18],"recognition":[19],"tasks.":[20],"In":[21],"this":[22],"paper,":[23],"we":[24,56,126,152],"analyze":[25,128],"the":[26,43,53,63,70,75,94,129,132,137,142,155,159,171],"GPU":[27,76],"performance":[28,77],"characteristics":[29,54,78],"of":[30,45,65,79,84,88,131,157],"five":[31],"popular":[32],"frameworks:":[35],"Caffe,":[36],"CNTK,":[37],"TensorFlow,":[38],"Theano,":[39],"and":[40,93,107,145],"Torch":[41],"perspective":[44],"a":[46,120],"representative":[47],"CNN":[48,66,112,133],"model,":[49],"AlexNet.":[50],"Based":[51],"on":[52,114],"obtained,":[55],"suggest":[57,99],"possible":[58],"optimization":[59],"methods":[60,108],"increase":[62,154],"efficiency":[64],"models":[67,113,134],"built":[68,135],"by":[69,136,165,170],"frameworks.":[71,172],"We":[72,97],"also":[73,98,127],"show":[74],"different":[80],"convolution":[81,103],"algorithms":[82,104],"each":[83],"which":[85],"uses":[86],"one":[87],"GEMM,":[89],"direct":[90],"convolution,":[91],"FFT,":[92],"Winograd":[95],"method.":[96],"criteria":[100],"choose":[102],"for":[105],"GPUs":[106],"build":[110],"efficient":[111],"GPUs.":[115],"Since":[116],"scaling":[117],"DNNs":[118],"multi-GPU":[121,143],"context":[122,144],"becomes":[123],"increasingly":[124],"important,":[125],"scalability":[130],"their":[146],"overhead.":[147],"The":[148],"result":[149],"indicates":[150],"can":[153],"speed":[156],"training":[158],"AlexNet":[160],"model":[161],"up":[162],"2X":[164],"just":[166],"changing":[167],"options":[168],"provided":[169]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":8},{"year":2024,"cited_by_count":14},{"year":2023,"cited_by_count":13},{"year":2022,"cited_by_count":13},{"year":2021,"cited_by_count":12},{"year":2020,"cited_by_count":16},{"year":2019,"cited_by_count":13},{"year":2018,"cited_by_count":18},{"year":2017,"cited_by_count":2}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}