{"id":"https://openalex.org/W7153998269","doi":"https://doi.org/10.1145/3805802","title":"PctoDL: Adaptive GPU Throughput Optimization for Deep Learning Inference with Power Constraints","display_name":"PctoDL: Adaptive GPU Throughput Optimization for Deep Learning Inference with Power Constraints","publication_year":2026,"publication_date":"2026-04-13","ids":{"openalex":"https://openalex.org/W7153998269","doi":"https://doi.org/10.1145/3805802"},"language":"en","primary_location":{"id":"doi:10.1145/3805802","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805802","pdf_url":null,"source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://doi.org/10.1145/3805802","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5060954730","display_name":"Meng Hao","orcid":"https://orcid.org/0000-0003-0043-4370"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Meng Hao","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0000-0003-0043-4370","affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133488539","display_name":"Zikun Wu","orcid":"https://orcid.org/0009-0003-8077-0047"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zikun Wu","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0009-0003-8077-0047","affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Xueyang Tian","orcid":"https://orcid.org/0009-0001-0415-327X"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xueyang Tian","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0009-0001-0415-327X","affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092178960","display_name":"Siyu Yang","orcid":null},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Siyu Yang","raw_affiliation_strings":["Computer Science and Technology, Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0009-0000-0071-6218","affiliations":[{"raw_affiliation_string":"Computer Science and Technology, Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133531300","display_name":"Guotong Guo","orcid":"https://orcid.org/0009-0004-0954-4001"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guotong Guo","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0009-0004-0954-4001","affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133516786","display_name":"Hongwei Yang","orcid":"https://orcid.org/0000-0002-8386-0131"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hongwei Yang","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0000-0002-8386-0131","affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133511496","display_name":"Hui He","orcid":"https://orcid.org/0000-0002-6494-775X"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hui He","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0000-0002-6494-775X","affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133507809","display_name":"Yiming Wang","orcid":"https://orcid.org/0000-0003-3298-5134"},"institutions":[{"id":"https://openalex.org/I4210142167","display_name":"DigitalSpace (United States)","ror":"https://ror.org/04h1th525","country_code":"US","type":"company","lineage":["https://openalex.org/I4210142167"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yiming Wang","raw_affiliation_strings":["School of Cyberspace Science, Harbin Industry University"],"raw_orcid":"https://orcid.org/0000-0003-3298-5134","affiliations":[{"raw_affiliation_string":"School of Cyberspace Science, Harbin Industry University","institution_ids":["https://openalex.org/I4210142167"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042525070","display_name":"Farui Wang","orcid":"https://orcid.org/0000-0002-2478-2153"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Farui Wang","raw_affiliation_strings":["Computer Science and Technology, Harbin Institute of Technology Shenzhen"],"raw_orcid":"https://orcid.org/0000-0002-2478-2153","affiliations":[{"raw_affiliation_string":"Computer Science and Technology, Harbin Institute of Technology Shenzhen","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123004106","display_name":"Desheng Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Desheng Wang","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0000-0002-7502-7094","affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5085221201","display_name":"Weizhe Zhang","orcid":"https://orcid.org/0000-0003-4783-876X"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Weizhe Zhang","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":"https://orcid.org/0000-0003-4783-876X","affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.465481,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"23","issue":"2","first_page":"1","last_page":"25"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.6899999976158142,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.6899999976158142,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.09589999914169312,"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.059700001031160355,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/inference","display_name":"Inference","score":0.6722999811172485},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.654699981212616},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.6444000005722046},{"id":"https://openalex.org/keywords/frequency-scaling","display_name":"Frequency scaling","score":0.619700014591217},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.47530001401901245},{"id":"https://openalex.org/keywords/power-optimization","display_name":"Power optimization","score":0.40400001406669617},{"id":"https://openalex.org/keywords/resource-allocation","display_name":"Resource allocation","score":0.40400001406669617},{"id":"https://openalex.org/keywords/greedy-algorithm","display_name":"Greedy algorithm","score":0.36000001430511475},{"id":"https://openalex.org/keywords/power-control","display_name":"Power control","score":0.34689998626708984}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8507999777793884},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6722999811172485},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.654699981212616},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6444000005722046},{"id":"https://openalex.org/C157742956","wikidata":"https://www.wikidata.org/wiki/Q3237776","display_name":"Frequency scaling","level":3,"score":0.619700014591217},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.47530001401901245},{"id":"https://openalex.org/C29202148","wikidata":"https://www.wikidata.org/wiki/Q287260","display_name":"Resource allocation","level":2,"score":0.40400001406669617},{"id":"https://openalex.org/C168292644","wikidata":"https://www.wikidata.org/wiki/Q10860336","display_name":"Power optimization","level":4,"score":0.40400001406669617},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.37869998812675476},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3628999888896942},{"id":"https://openalex.org/C51823790","wikidata":"https://www.wikidata.org/wiki/Q504353","display_name":"Greedy algorithm","level":2,"score":0.36000001430511475},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3589000105857849},{"id":"https://openalex.org/C56685638","wikidata":"https://www.wikidata.org/wiki/Q2300474","display_name":"Power control","level":3,"score":0.34689998626708984},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.3431999981403351},{"id":"https://openalex.org/C2780609101","wikidata":"https://www.wikidata.org/wiki/Q17156588","display_name":"Resource management (computing)","level":2,"score":0.3310999870300293},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.3239000141620636},{"id":"https://openalex.org/C203479927","wikidata":"https://www.wikidata.org/wiki/Q5165939","display_name":"Controller (irrigation)","level":2,"score":0.3133000135421753},{"id":"https://openalex.org/C14961307","wikidata":"https://www.wikidata.org/wiki/Q5377176","display_name":"Energy minimization","level":2,"score":0.3109999895095825},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.3095000088214874},{"id":"https://openalex.org/C2778774385","wikidata":"https://www.wikidata.org/wiki/Q4437810","display_name":"Power management","level":3,"score":0.3077999949455261},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.305400013923645},{"id":"https://openalex.org/C45872418","wikidata":"https://www.wikidata.org/wiki/Q5318966","display_name":"Dynamic demand","level":3,"score":0.305400013923645},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.29409998655319214},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2930000126361847},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.2921000123023987},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.29109999537467957},{"id":"https://openalex.org/C163258240","wikidata":"https://www.wikidata.org/wiki/Q25342","display_name":"Power (physics)","level":2,"score":0.28349998593330383},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.2728999853134155},{"id":"https://openalex.org/C2780165032","wikidata":"https://www.wikidata.org/wiki/Q16869822","display_name":"Energy consumption","level":2,"score":0.25450000166893005}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3805802","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805802","pdf_url":null,"source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1145/3805802","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805802","pdf_url":null,"source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"},"sustainable_development_goals":[{"score":0.8235412240028381,"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7"}],"awards":[{"id":"https://openalex.org/G2686570925","display_name":null,"funder_award_id":"2025KF1A0020","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G3095772617","display_name":null,"funder_award_id":"62202123, and 62402141","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G752701465","display_name":null,"funder_award_id":"U22A2036","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":27,"referenced_works":["https://openalex.org/W2064707793","https://openalex.org/W2108598243","https://openalex.org/W2119112357","https://openalex.org/W2194775991","https://openalex.org/W2410158158","https://openalex.org/W2612225380","https://openalex.org/W2613168994","https://openalex.org/W2794670651","https://openalex.org/W2945148926","https://openalex.org/W2953736860","https://openalex.org/W2963163009","https://openalex.org/W2963446712","https://openalex.org/W2963918968","https://openalex.org/W2998957070","https://openalex.org/W3097411828","https://openalex.org/W3116103263","https://openalex.org/W3166036439","https://openalex.org/W4206149382","https://openalex.org/W4206943884","https://openalex.org/W4286635279","https://openalex.org/W4312847199","https://openalex.org/W4313229743","https://openalex.org/W4386703510","https://openalex.org/W4388116548","https://openalex.org/W4390968511","https://openalex.org/W4394944658","https://openalex.org/W7123411730"],"related_works":[],"abstract_inverted_index":{"The":[0],"proliferation":[1],"of":[2,136,148,154,166],"deep":[3],"learning":[4],"inference":[5,56],"services":[6],"in":[7,75],"power-constrained":[8,76],"environments":[9],"necessitates":[10],"GPU":[11],"management":[12],"strategies":[13],"that":[14,53],"maximize":[15],"throughput":[16,57,127],"within":[17],"strict":[18],"power":[19,90],"envelopes.":[20],"Existing":[21],"approaches":[22],"often":[23],"treat":[24],"frequency":[25,68,85,106],"scaling":[26],"and":[27,66,108,150,173],"resource":[28,62,81],"partitioning":[29,82,98,160],"as":[30],"orthogonal":[31],"problems":[32],"or":[33],"rely":[34],"on":[35,168,175],"static":[36],"hardware":[37],"assumptions,":[38],"leading":[39],"to":[40],"suboptimal":[41],"energy":[42],"efficiency.":[43],"This":[44],"article":[45],"presents":[46],"PctoDL":[47,79,124,162],",":[48],"a":[49,88,94,100,133,151],"power-aware":[50],"scheduling":[51],"system":[52],"maximizes":[54],"aggregate":[55],"by":[58,130],"jointly":[59],"optimizing":[60],"spatial":[61],"partitioning,":[63],"batch":[64],"size,":[65],"SM/memory":[67],"settings.":[69],"To":[70],"address":[71],"the":[72,118,139,169,176],"throughput\u2013power":[73],"tradeoff":[74],"multi-tenant":[77],"inference,":[78],"couples":[80],"with":[83,132,157],"coordinated":[84],"control":[86],"under":[87],"fixed":[89],"cap.":[91],"It":[92],"combines":[93],"physics-informed":[95],"iterative":[96],"greedy":[97],"algorithm,":[99],"thermodynamic":[101],"model-predictive":[102],"controller":[103],"for":[104,114],"runtime":[105],"regulation,":[107],"an":[109,145],"online":[110],"joint":[111],"optimization":[112],"mechanism":[113],"adaptive":[115],"refinement.":[116],"On":[117,138],"NVIDIA":[119,140],"RTX":[120,170],"3080":[121,171],"Ti":[122,172],"platform,":[123,142],"improves":[125],"average":[126,146],"over":[128],"BatchDVFS":[129],"108.41%,":[131],"peak":[134],"gain":[135,147,153],"262.74%.":[137],"A100":[141],"it":[143],"delivers":[144],"19.74%":[149],"maximum":[152],"57.03%.":[155],"Compared":[156],"Morak\u2019s":[158],"coarse-grained":[159],"approach,":[161],"achieves":[163],"average/peak":[164],"gains":[165],"79.05%/137.93%":[167],"26.33%/70.21%":[174],"A100.":[177]},"counts_by_year":[],"updated_date":"2026-06-26T06:17:10.115597","created_date":"2026-04-14T00:00:00"}