{"id":"https://openalex.org/W3136752701","doi":"https://doi.org/10.1109/bigdata50022.2020.9378252","title":"A Topology-Aware Performance Prediction Model for Distributed Deep Learning on GPU Clusters","display_name":"A Topology-Aware Performance Prediction Model for Distributed Deep Learning on GPU Clusters","publication_year":2020,"publication_date":"2020-12-10","ids":{"openalex":"https://openalex.org/W3136752701","doi":"https://doi.org/10.1109/bigdata50022.2020.9378252","mag":"3136752701"},"language":"en","primary_location":{"id":"doi:10.1109/bigdata50022.2020.9378252","is_oa":false,"landing_page_url":"https://doi.org/10.1109/bigdata50022.2020.9378252","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 IEEE International Conference on Big Data (Big Data)","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/A5020544285","display_name":"Zheyu Lin","orcid":null},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Zheyu Lin","raw_affiliation_strings":["School of ECE, Peking University"],"affiliations":[{"raw_affiliation_string":"School of ECE, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019232463","display_name":"Xukun Chen","orcid":null},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xukun Chen","raw_affiliation_strings":["Computer Science Department, Peking University"],"affiliations":[{"raw_affiliation_string":"Computer Science Department, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101987114","display_name":"Hanyu Zhao","orcid":"https://orcid.org/0000-0002-2536-0016"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hanyu Zhao","raw_affiliation_strings":["Computer Science Department, Peking University"],"affiliations":[{"raw_affiliation_string":"Computer Science Department, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009149646","display_name":"Yunteng Luan","orcid":null},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yunteng Luan","raw_affiliation_strings":["Computer Science Department, Peking University"],"affiliations":[{"raw_affiliation_string":"Computer Science Department, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102862758","display_name":"Zhi Yang","orcid":"https://orcid.org/0000-0002-8219-4499"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhi Yang","raw_affiliation_strings":["Computer Science Department, Peking University"],"affiliations":[{"raw_affiliation_string":"Computer Science Department, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5103039363","display_name":"Yafei Dai","orcid":"https://orcid.org/0000-0003-2126-8607"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yafei Dai","raw_affiliation_strings":["Computer Science Department, Peking University"],"affiliations":[{"raw_affiliation_string":"Computer Science Department, Peking University","institution_ids":["https://openalex.org/I20231570"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5020544285"],"corresponding_institution_ids":["https://openalex.org/I20231570"],"apc_list":null,"apc_paid":null,"fwci":0.8007,"has_fulltext":false,"cited_by_count":6,"citation_normalized_percentile":{"value":0.82062233,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"2795","last_page":"2801"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10101","display_name":"Cloud Computing and Resource Management","score":0.998199999332428,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10101","display_name":"Cloud Computing and Resource Management","score":0.998199999332428,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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":0.9976999759674072,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9962000250816345,"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":0.8253765106201172},{"id":"https://openalex.org/keywords/locality","display_name":"Locality","score":0.7046263813972473},{"id":"https://openalex.org/keywords/gpu-cluster","display_name":"GPU cluster","score":0.6742473840713501},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.5770876407623291},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5282860398292542},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.49667078256607056},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4918792247772217},{"id":"https://openalex.org/keywords/job-scheduler","display_name":"Job scheduler","score":0.45844268798828125},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.45058107376098633},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.37112128734588623},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.16972616314888},{"id":"https://openalex.org/keywords/mathematical-optimization","display_name":"Mathematical optimization","score":0.10594320297241211},{"id":"https://openalex.org/keywords/cloud-computing","display_name":"Cloud computing","score":0.10334041714668274}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8253765106201172},{"id":"https://openalex.org/C2779808786","wikidata":"https://www.wikidata.org/wiki/Q6664603","display_name":"Locality","level":2,"score":0.7046263813972473},{"id":"https://openalex.org/C2781335571","wikidata":"https://www.wikidata.org/wiki/Q2633544","display_name":"GPU cluster","level":3,"score":0.6742473840713501},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.5770876407623291},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5282860398292542},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.49667078256607056},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4918792247772217},{"id":"https://openalex.org/C111873713","wikidata":"https://www.wikidata.org/wiki/Q1641413","display_name":"Job scheduler","level":3,"score":0.45844268798828125},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.45058107376098633},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.37112128734588623},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.16972616314888},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.10594320297241211},{"id":"https://openalex.org/C79974875","wikidata":"https://www.wikidata.org/wiki/Q483639","display_name":"Cloud computing","level":2,"score":0.10334041714668274},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0},{"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/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/bigdata50022.2020.9378252","is_oa":false,"landing_page_url":"https://doi.org/10.1109/bigdata50022.2020.9378252","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 IEEE International Conference on Big Data (Big Data)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.5400000214576721,"display_name":"Decent work and economic growth","id":"https://metadata.un.org/sdg/8"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":29,"referenced_works":["https://openalex.org/W1686810756","https://openalex.org/W1922655562","https://openalex.org/W2018162709","https://openalex.org/W2060393849","https://openalex.org/W2083842231","https://openalex.org/W2102709380","https://openalex.org/W2137028279","https://openalex.org/W2186615578","https://openalex.org/W2194775991","https://openalex.org/W2309679942","https://openalex.org/W2402144811","https://openalex.org/W2546571074","https://openalex.org/W2752512710","https://openalex.org/W2798515322","https://openalex.org/W2953384591","https://openalex.org/W2963403751","https://openalex.org/W2963403868","https://openalex.org/W3095922049","https://openalex.org/W4243267510","https://openalex.org/W4253824360","https://openalex.org/W4385245566","https://openalex.org/W6637373629","https://openalex.org/W6640090968","https://openalex.org/W6686509673","https://openalex.org/W6697698479","https://openalex.org/W6713134421","https://openalex.org/W6739901393","https://openalex.org/W6744307745","https://openalex.org/W6758283263"],"related_works":["https://openalex.org/W1555349535","https://openalex.org/W1556451512","https://openalex.org/W4234091740","https://openalex.org/W4213350282","https://openalex.org/W2230171082","https://openalex.org/W2583128298","https://openalex.org/W2014963843","https://openalex.org/W44553394","https://openalex.org/W4386430090","https://openalex.org/W2755738839"],"abstract_inverted_index":{"Today,":[0],"multi-GPU":[1],"training":[2,16,70,80],"has":[3],"become":[4],"a":[5,15,83,98,123,146,150,159],"common":[6],"practice":[7],"for":[8,49,67,97],"deep":[9,68],"learning":[10,69],"workloads.":[11],"The":[12,101],"performance":[13,47,128,134],"of":[14,35,42,45,85,95,108,116,126],"job":[17,55,165],"could":[18],"be":[19],"affected":[20],"significantly":[21],"by":[22,168,172],"both":[23,54],"the":[24,28,32,36,40,43,79,91,109,113],"GPU":[25,73],"connectivity":[26],"in":[27],"system":[29,110],"topology":[30,111],"and":[31,56,93,112,170],"computation-communication":[33],"pattern":[34],"job.":[37,100],"This":[38,75],"highlights":[39],"necessity":[41],"awareness":[44],"jobs\u2019":[46],"characteristics":[48],"cluster":[50,57],"schedulers":[51],"to":[52,143],"improve":[53],"efficiency.In":[58],"this":[59],"paper,":[60],"we":[61],"propose":[62],"an":[63],"online":[64,120],"resource-performance":[65],"model":[66,76,102,135,157],"jobs":[71,118],"on":[72,105,122,140,149],"clusters.":[74],"can":[77],"estimate":[78],"speed":[81],"as":[82],"function":[84],"any":[86],"given":[87],"resource":[88],"setting":[89],"(i.e.,":[90],"number":[92],"locality":[94],"GPUs)":[96],"specific":[99],"is":[103],"based":[104],"systematic":[106],"modeling":[107],"communication":[114],"patterns":[115],"individual":[117],"with":[119],"fitting":[121],"sample":[124],"set":[125],"profiled":[127],"data.":[129],"Experiments":[130],"show":[131],"that":[132,155],"our":[133,156],"achieves":[136],"94%":[137],"prediction":[138],"accuracy":[139],"average":[141,164],"(up":[142],"99.9%).":[144],"Additionally,":[145],"large-scale":[147],"simulation":[148],"real":[151],"production":[152],"trace":[153],"demonstrates":[154],"helps":[158],"typical":[160],"scheduling":[161],"algorithm":[162],"decrease":[163],"completion":[166],"time":[167],"3.4x":[169],"makespan":[171],"1.7x.":[173]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":1},{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":2},{"year":2022,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}