{"id":"https://openalex.org/W4386707702","doi":"https://doi.org/10.1145/3605573.3605603","title":"Mercury: Fast and Optimal Device Placement for Large Deep Learning Models","display_name":"Mercury: Fast and Optimal Device Placement for Large Deep Learning Models","publication_year":2023,"publication_date":"2023-08-07","ids":{"openalex":"https://openalex.org/W4386707702","doi":"https://doi.org/10.1145/3605573.3605603"},"language":"en","primary_location":{"id":"doi:10.1145/3605573.3605603","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3605573.3605603","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3605573.3605603","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 52nd International Conference on Parallel Processing","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3605573.3605603","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101404511","display_name":"Hengwei Xu","orcid":"https://orcid.org/0000-0002-1363-9394"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hengwei Xu","raw_affiliation_strings":["University Of Science And Technology Of China, China"],"raw_orcid":"https://orcid.org/0000-0002-1363-9394","affiliations":[{"raw_affiliation_string":"University Of Science And Technology Of China, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004288865","display_name":"Pengyuan Zhou","orcid":"https://orcid.org/0000-0002-7909-4059"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Pengyuan Zhou","raw_affiliation_strings":["University of Science and Technology of China, China"],"raw_orcid":"https://orcid.org/0000-0002-7909-4059","affiliations":[{"raw_affiliation_string":"University of Science and Technology of China, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5081885053","display_name":"Haiyong Xie","orcid":"https://orcid.org/0000-0003-2084-2697"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Haiyong Xie","raw_affiliation_strings":["Advanced Innovation Center for Human Brain Protection, capital medical university, China"],"raw_orcid":"https://orcid.org/0000-0003-2084-2697","affiliations":[{"raw_affiliation_string":"Advanced Innovation Center for Human Brain Protection, capital medical university, China","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5051983841","display_name":"Yong Liao","orcid":"https://orcid.org/0000-0001-6403-0557"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yong Liao","raw_affiliation_strings":["University of Science and Technology of China, China"],"raw_orcid":"https://orcid.org/0000-0001-6403-0557","affiliations":[{"raw_affiliation_string":"University of Science and Technology of China, China","institution_ids":["https://openalex.org/I126520041"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.1213,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.42043564,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":null,"issue":null,"first_page":"412","last_page":"422"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9997000098228455,"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":0.9997000098228455,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.996399998664856,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9955999851226807,"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/computer-science","display_name":"Computer science","score":0.8060042262077332},{"id":"https://openalex.org/keywords/mercury","display_name":"Mercury (programming language)","score":0.7811139822006226},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.6950399875640869},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.45046159625053406},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.4374474883079529},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.4027625620365143},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.30472278594970703},{"id":"https://openalex.org/keywords/mathematical-optimization","display_name":"Mathematical optimization","score":0.10634052753448486}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8060042262077332},{"id":"https://openalex.org/C2777777548","wikidata":"https://www.wikidata.org/wiki/Q206040","display_name":"Mercury (programming language)","level":2,"score":0.7811139822006226},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6950399875640869},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.45046159625053406},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4374474883079529},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.4027625620365143},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.30472278594970703},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.10634052753448486},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","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":2,"locations":[{"id":"doi:10.1145/3605573.3605603","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3605573.3605603","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3605573.3605603","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 52nd International Conference on Parallel Processing","raw_type":"proceedings-article"},{"id":"pmh:oai:pure.atira.dk:publications/49e3b5a5-b117-4f30-8c5b-50a159935fde","is_oa":false,"landing_page_url":"https://pure.au.dk/portal/en/publications/49e3b5a5-b117-4f30-8c5b-50a159935fde","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Xu , H , Zhou , P , Xie , H & Liao , Y 2023 , Mercury : Fast and Optimal Device Placement for Large Deep Learning Models . in 52nd International Conference on Parallel Processing, ICPP 2023 - Main Conference Proceedings . Association for Computing Machinery, Inc. , pp. 412-422 , 52nd International Conference on Parallel Processing, ICPP 2023 , Salt Lake City , United States , 07/08/2023 . https://doi.org/10.1145/3605573.3605603","raw_type":"contributionToPeriodical"}],"best_oa_location":{"id":"doi:10.1145/3605573.3605603","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3605573.3605603","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3605573.3605603","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 52nd International Conference on Parallel Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.4099999964237213,"display_name":"No poverty","id":"https://metadata.un.org/sdg/1"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4386707702.pdf","grobid_xml":"https://content.openalex.org/works/W4386707702.grobid-xml"},"referenced_works_count":20,"referenced_works":["https://openalex.org/W2002292054","https://openalex.org/W2097117768","https://openalex.org/W2169878774","https://openalex.org/W2183341477","https://openalex.org/W2588488651","https://openalex.org/W2792063935","https://openalex.org/W3035078899","https://openalex.org/W3039050620","https://openalex.org/W3047687953","https://openalex.org/W3101965820","https://openalex.org/W3104001151","https://openalex.org/W3109215866","https://openalex.org/W3112483307","https://openalex.org/W3138516171","https://openalex.org/W3177451412","https://openalex.org/W3195905284","https://openalex.org/W3202656791","https://openalex.org/W4287750421","https://openalex.org/W6713134421","https://openalex.org/W6784370339"],"related_works":["https://openalex.org/W2050515154","https://openalex.org/W2266983138","https://openalex.org/W1990269663","https://openalex.org/W2066916115","https://openalex.org/W2317577920","https://openalex.org/W2067648061","https://openalex.org/W2026418612","https://openalex.org/W3005719954","https://openalex.org/W2063552428","https://openalex.org/W2082425661"],"abstract_inverted_index":{"The":[0,91],"rapidly":[1],"expanding":[2],"neural":[3],"network":[4],"models":[5],"are":[6],"becoming":[7],"increasingly":[8],"challenging":[9],"to":[10,24,82,115],"run":[11],"on":[12,87],"a":[13,46,58,74],"single":[14],"device.":[15],"Hence,":[16],"model":[17,62,110],"parallelism":[18],"over":[19],"multiple":[20],"devices":[21],"is":[22],"critical":[23],"guaranteeing":[25],"the":[26,66,99,109,116],"efficiency":[27],"of":[28,76],"training":[29],"large":[30,54,89],"models.":[31,55,90],"Recent":[32],"proposals":[33],"either":[34],"have":[35],"long":[36],"processing":[37],"time":[38,103],"or":[39],"poor":[40],"performance.":[41],"Therefore,":[42],"we":[43],"propose":[44],"Mercury,":[45],"fast":[47],"framework":[48],"for":[49,53],"optimizing":[50],"device":[51],"placement":[52,71,100],"Mercury":[56,86,95],"employs":[57],"simple":[59],"but":[60,106],"efficient":[61],"parallelization":[63],"strategy":[64],"in":[65],"baseline":[67],"measurement,":[68],"and":[69,84],"generates":[70],"policies":[72],"through":[73],"series":[75],"scheduling":[77],"algorithms.":[78],"We":[79],"conduct":[80],"experiments":[81],"deploy":[83],"evaluate":[85],"numerous":[88],"results":[92],"show":[93],"that":[94],"not":[96],"only":[97],"reduces":[98],"policy":[101],"generation":[102],"by":[104,112],"26.4%":[105],"also":[107],"improves":[108],"throughput":[111],"218.5%":[113],"compared":[114],"most":[117],"advanced":[118],"methods.":[119]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-08-01T09:00:35.917206","created_date":"2025-10-10T00:00:00"}