{"id":"https://openalex.org/W2962711501","doi":"https://doi.org/10.1145/3337821.3337839","title":"A Unified Optimization Approach for CNN Model Inference on Integrated GPUs","display_name":"A Unified Optimization Approach for CNN Model Inference on Integrated GPUs","publication_year":2019,"publication_date":"2019-07-25","ids":{"openalex":"https://openalex.org/W2962711501","doi":"https://doi.org/10.1145/3337821.3337839","mag":"2962711501"},"language":"en","primary_location":{"id":"doi:10.1145/3337821.3337839","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3337821.3337839","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3337821.3337839","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 48th International Conference on Parallel Processing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3337821.3337839","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5070031492","display_name":"Leyuan Wang","orcid":"https://orcid.org/0000-0003-4084-4470"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Leyuan Wang","raw_affiliation_strings":["Amazon Web Services, East Palo Alto, CA, USA"],"affiliations":[{"raw_affiliation_string":"Amazon Web Services, East Palo Alto, CA, USA","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100456835","display_name":"Zhi Chen","orcid":"https://orcid.org/0000-0002-9385-144X"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhi Chen","raw_affiliation_strings":["Amazon Web Services, East Palo Alto, CA, USA"],"affiliations":[{"raw_affiliation_string":"Amazon Web Services, East Palo Alto, CA, USA","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100701317","display_name":"Yizhi Liu","orcid":"https://orcid.org/0000-0002-2067-2707"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yizhi Liu","raw_affiliation_strings":["Amazon Web Services, East Palo Alto, CA, USA"],"affiliations":[{"raw_affiliation_string":"Amazon Web Services, East Palo Alto, CA, USA","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100319039","display_name":"Yao Wang","orcid":"https://orcid.org/0000-0003-3199-3802"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yao Wang","raw_affiliation_strings":["Amazon Web Services East Palo Alto, CA, USA","Amazon Web Services, East Palo Alto, CA, USA"],"affiliations":[{"raw_affiliation_string":"Amazon Web Services East Palo Alto, CA, USA","institution_ids":["https://openalex.org/I1311688040"]},{"raw_affiliation_string":"Amazon Web Services, East Palo Alto, CA, USA","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5061339425","display_name":"Lianmin Zheng","orcid":"https://orcid.org/0000-0001-5812-731X"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lianmin Zheng","raw_affiliation_strings":["Shanghai Jiaotong University, Shanghai, China"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiaotong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100399461","display_name":"Mu Li","orcid":"https://orcid.org/0000-0002-4433-2301"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mu Li","raw_affiliation_strings":["Amazon Web Services, East Palo Alto, CA, USA"],"affiliations":[{"raw_affiliation_string":"Amazon Web Services, East Palo Alto, CA, USA","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101638214","display_name":"Yida Wang","orcid":"https://orcid.org/0000-0001-8165-840X"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yida Wang","raw_affiliation_strings":["Amazon Web Services East Palo Alto, CA, USA","Amazon Web Services, East Palo Alto, CA, USA"],"affiliations":[{"raw_affiliation_string":"Amazon Web Services East Palo Alto, CA, USA","institution_ids":["https://openalex.org/I1311688040"]},{"raw_affiliation_string":"Amazon Web Services, East Palo Alto, CA, USA","institution_ids":["https://openalex.org/I1311688040"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5070031492"],"corresponding_institution_ids":["https://openalex.org/I1311688040"],"apc_list":null,"apc_paid":null,"fwci":2.4507,"has_fulltext":true,"cited_by_count":38,"citation_normalized_percentile":{"value":0.91530046,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"10"},"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/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9969000220298767,"subfield":{"id":"https://openalex.org/subfields/2808","display_name":"Neurology"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9929999709129333,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8571851253509521},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6423413157463074},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.5264198184013367},{"id":"https://openalex.org/keywords/graphics","display_name":"Graphics","score":0.5223230719566345},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.5184668898582458},{"id":"https://openalex.org/keywords/cloud-computing","display_name":"Cloud computing","score":0.5014815330505371},{"id":"https://openalex.org/keywords/programming-paradigm","display_name":"Programming paradigm","score":0.49898576736450195},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.4768996238708496},{"id":"https://openalex.org/keywords/general-purpose-computing-on-graphics-processing-units","display_name":"General-purpose computing on graphics processing units","score":0.4655132293701172},{"id":"https://openalex.org/keywords/edge-device","display_name":"Edge device","score":0.4554164409637451},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.43078354001045227},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3733944296836853},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.35290274024009705},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.1570444107055664},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.10312491655349731}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8571851253509521},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6423413157463074},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.5264198184013367},{"id":"https://openalex.org/C21442007","wikidata":"https://www.wikidata.org/wiki/Q1027879","display_name":"Graphics","level":2,"score":0.5223230719566345},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.5184668898582458},{"id":"https://openalex.org/C79974875","wikidata":"https://www.wikidata.org/wiki/Q483639","display_name":"Cloud computing","level":2,"score":0.5014815330505371},{"id":"https://openalex.org/C34165917","wikidata":"https://www.wikidata.org/wiki/Q188267","display_name":"Programming paradigm","level":2,"score":0.49898576736450195},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.4768996238708496},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.4655132293701172},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.4554164409637451},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.43078354001045227},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3733944296836853},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.35290274024009705},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.1570444107055664},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.10312491655349731},{"id":"https://openalex.org/C21547014","wikidata":"https://www.wikidata.org/wiki/Q1423657","display_name":"Operations management","level":1,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3337821.3337839","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3337821.3337839","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3337821.3337839","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 48th International Conference on Parallel Processing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3337821.3337839","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3337821.3337839","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3337821.3337839","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 48th International Conference on Parallel Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","score":0.8899999856948853,"display_name":"Affordable and clean energy"}],"awards":[{"id":"https://openalex.org/G401493834","display_name":null,"funder_award_id":"FA8650-18-2-7836","funder_id":"https://openalex.org/F4320338294","funder_display_name":"Air Force Research Laboratory"},{"id":"https://openalex.org/G4795485331","display_name":null,"funder_award_id":"FA8650-18-2-7836","funder_id":"https://openalex.org/F4320332180","funder_display_name":"Defense Advanced Research Projects Agency"}],"funders":[{"id":"https://openalex.org/F4320332180","display_name":"Defense Advanced Research Projects Agency","ror":"https://ror.org/02caytj08"},{"id":"https://openalex.org/F4320332815","display_name":"Advanced Research Projects Agency","ror":"https://ror.org/02caytj08"},{"id":"https://openalex.org/F4320338294","display_name":"Air Force Research Laboratory","ror":"https://ror.org/02e2egq70"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W2962711501.pdf","grobid_xml":"https://content.openalex.org/works/W2962711501.grobid-xml"},"referenced_works_count":38,"referenced_works":["https://openalex.org/W1678662003","https://openalex.org/W1967810725","https://openalex.org/W2055312318","https://openalex.org/W2079248286","https://openalex.org/W2119144962","https://openalex.org/W2119547137","https://openalex.org/W2142805836","https://openalex.org/W2161061943","https://openalex.org/W2193145675","https://openalex.org/W2194775991","https://openalex.org/W2279098554","https://openalex.org/W2297325673","https://openalex.org/W2335962142","https://openalex.org/W2544002786","https://openalex.org/W2582737986","https://openalex.org/W2612445135","https://openalex.org/W2620106252","https://openalex.org/W2626129225","https://openalex.org/W2725001169","https://openalex.org/W2734941459","https://openalex.org/W2796347433","https://openalex.org/W2799002811","https://openalex.org/W2804032941","https://openalex.org/W2898537641","https://openalex.org/W2904295992","https://openalex.org/W2931700347","https://openalex.org/W2931743911","https://openalex.org/W2963125010","https://openalex.org/W2963947383","https://openalex.org/W2990138404","https://openalex.org/W3101616732","https://openalex.org/W3105314253","https://openalex.org/W3106250896","https://openalex.org/W4250981202","https://openalex.org/W4251637954","https://openalex.org/W6683923952","https://openalex.org/W6752057402","https://openalex.org/W6754777574"],"related_works":["https://openalex.org/W2319468889","https://openalex.org/W2319727539","https://openalex.org/W2080914619","https://openalex.org/W2794923745","https://openalex.org/W2187537688","https://openalex.org/W2498443814","https://openalex.org/W2887103494","https://openalex.org/W2187267005","https://openalex.org/W2121341638","https://openalex.org/W2045275492"],"abstract_inverted_index":{"Modern":[0],"deep":[1],"learning":[2],"applications":[3],"urge":[4],"to":[5,31,67,102,117,134,155,174,184,213,240],"push":[6],"the":[7,13,25,28,32,46,53,56,61,71,75,90,103,123,127,190],"model":[8,50,120,234],"inference":[9,121],"taking":[10],"place":[11],"at":[12,74,126],"edge":[14,76],"devices":[15],"for":[16,168],"multiple":[17,144],"reasons":[18],"such":[19],"as":[20,146,148],"achieving":[21],"shorter":[22],"latency,":[23],"relieving":[24],"burden":[26],"of":[27,45,60,105,219],"network":[29],"connecting":[30],"cloud,":[33],"and":[34,81,85,108,136,200,223,236,254],"protecting":[35],"user":[36],"privacy.":[37],"The":[38,178],"Convolutional":[39],"Neural":[40],"Network":[41],"(CNN)":[42],"is":[43,65,99,237,255],"one":[44],"most":[47],"widely":[48],"used":[49],"family":[51],"in":[52,249,252],"applications.":[54],"Given":[55],"high":[57],"computational":[58],"complexity":[59],"CNN":[62,119],"models,":[63],"it":[64],"favorable":[66],"execute":[68,118],"them":[69],"on":[70,95,122,140,176,194,216],"integrated":[72,96,124,141,202],"GPUs":[73,97,125,142],"devices,":[77],"which":[78,129],"are":[79],"ubiquitous":[80],"have":[82],"more":[83,238],"power":[84],"better":[86,211],"energy":[87],"efficiency":[88],"than":[89],"accompanying":[91],"CPUs.":[92],"However,":[93],"programming":[94,109],"efficiently":[98],"challenging":[100],"due":[101],"variety":[104],"their":[106],"architectures":[107],"interfaces.":[110],"This":[111],"paper":[112],"proposes":[113],"an":[114],"end-to-end":[115],"solution":[116,162,206,230,245],"edge,":[128],"uses":[130],"a":[131,165,217,232],"unified":[132],"IR":[133],"represent":[135],"optimize":[137,156],"vision-specific":[138],"operators":[139,158,169],"from":[143],"vendors,":[145],"well":[147],"leverages":[149],"machine":[150],"learning-based":[151],"scheduling":[152],"search":[153],"schemes":[154],"computationally-intensive":[157],"like":[159],"convolution.":[160],"Our":[161,244],"even":[163,210],"provides":[164],"fallback":[166],"mechanism":[167],"not":[170],"suitable":[171],"or":[172,209],"convenient":[173],"run":[175],"GPUs.":[177],"evaluation":[179],"results":[180],"suggest":[181],"that":[182],"compared":[183],"state-of-the-art":[185],"solutions":[186],"backed":[187],"up":[188],"by":[189],"vendor-provided":[191],"high-performance":[192],"libraries":[193],"Intel":[195],"Graphics,":[196],"ARM":[197],"Mali":[198],"GPU,":[199,204],"Nvidia":[201],"Maxwell":[203],"our":[205,229],"achieves":[207],"similar,":[208],"(up":[212],"1.62\u00d7),":[214],"performance":[215],"number":[218],"popular":[220],"image":[221],"classification":[222],"object":[224],"detection":[225],"models.":[226,243],"In":[227],"addition,":[228],"has":[231,246],"wider":[233],"coverage":[235],"flexible":[239],"embrace":[241],"new":[242],"been":[247],"adopted":[248],"production":[250],"services":[251],"AWS":[253],"open-sourced.":[256]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":4},{"year":2023,"cited_by_count":3},{"year":2022,"cited_by_count":6},{"year":2021,"cited_by_count":4},{"year":2020,"cited_by_count":13},{"year":2019,"cited_by_count":1}],"updated_date":"2026-04-24T08:23:43.765630","created_date":"2025-10-10T00:00:00"}