{"id":"https://openalex.org/W4404795623","doi":"https://doi.org/10.1109/tpds.2024.3507814","title":"Towards Universal Performance Modeling for Machine Learning Training on Multi-GPU Platforms","display_name":"Towards Universal Performance Modeling for Machine Learning Training on Multi-GPU Platforms","publication_year":2024,"publication_date":"2024-11-28","ids":{"openalex":"https://openalex.org/W4404795623","doi":"https://doi.org/10.1109/tpds.2024.3507814"},"language":"en","primary_location":{"id":"doi:10.1109/tpds.2024.3507814","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2024.3507814","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Parallel and Distributed Systems","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://escholarship.org/uc/item/35n7r10r","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5082748769","display_name":"Zhongyi Lin","orcid":"https://orcid.org/0000-0002-5992-3913"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]},{"id":"https://openalex.org/I4210137977","display_name":"Advanced Micro Devices (United States)","ror":"https://ror.org/04kd6c783","country_code":"US","type":"company","lineage":["https://openalex.org/I4210137977"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhongyi Lin","raw_affiliation_strings":["Advanced Micro Devices, Inc, San Jose, CA, USA","Meta, Inc, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-5992-3913","affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc, San Jose, CA, USA","institution_ids":["https://openalex.org/I4210137977"]},{"raw_affiliation_string":"Meta, Inc, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041345011","display_name":"Ning Sun","orcid":"https://orcid.org/0000-0002-6907-3756"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ning Sun","raw_affiliation_strings":["Meta, Inc, Menlo Park, CA, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta, Inc, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Pallab Bhattacharya","orcid":"https://orcid.org/0009-0005-7653-744X"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Pallab Bhattacharya","raw_affiliation_strings":["NVIDIA, Santa Clara, CA, USA"],"raw_orcid":"https://orcid.org/0009-0005-7653-744X","affiliations":[{"raw_affiliation_string":"NVIDIA, Santa Clara, CA, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101060608","display_name":"Xizhou Feng","orcid":null},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xizhou Feng","raw_affiliation_strings":["Meta, Inc, Menlo Park, CA, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta, Inc, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100808706","display_name":"Louis Feng","orcid":null},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Louis Feng","raw_affiliation_strings":["Meta, Inc, Menlo Park, CA, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta, Inc, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5028662746","display_name":"John D. Owens","orcid":"https://orcid.org/0000-0001-6582-8237"},"institutions":[{"id":"https://openalex.org/I84218800","display_name":"University of California, Davis","ror":"https://ror.org/05rrcem69","country_code":"US","type":"education","lineage":["https://openalex.org/I84218800"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"John D. Owens","raw_affiliation_strings":["Department of Electrical and Computer Engineering, University of California, Davis, CA, USA"],"raw_orcid":"https://orcid.org/0000-0001-6582-8237","affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering, University of California, Davis, CA, USA","institution_ids":["https://openalex.org/I84218800"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.0686,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.9053905,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":"36","issue":"2","first_page":"226","last_page":"238"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10101","display_name":"Cloud Computing and Resource Management","score":0.9553999900817871,"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":0.9553999900817871,"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/T11122","display_name":"Online Learning and Analytics","score":0.9455000162124634,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.9276000261306763,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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":0.8687015771865845},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.5460978150367737},{"id":"https://openalex.org/keywords/general-purpose-computing-on-graphics-processing-units","display_name":"General-purpose computing on graphics processing units","score":0.4935241937637329},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.47625732421875},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.4649601876735687},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.36212265491485596},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.35967400670051575},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.33911317586898804},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.33374661207199097},{"id":"https://openalex.org/keywords/graphics","display_name":"Graphics","score":0.2060094177722931}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8687015771865845},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5460978150367737},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.4935241937637329},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.47625732421875},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.4649601876735687},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.36212265491485596},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.35967400670051575},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.33911317586898804},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.33374661207199097},{"id":"https://openalex.org/C21442007","wikidata":"https://www.wikidata.org/wiki/Q1027879","display_name":"Graphics","level":2,"score":0.2060094177722931},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/tpds.2024.3507814","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2024.3507814","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Parallel and Distributed Systems","raw_type":"journal-article"},{"id":"pmh:oai:escholarship.org:ark:/13030/qt35n7r10r","is_oa":true,"landing_page_url":"https://escholarship.org/uc/item/35n7r10r","pdf_url":null,"source":{"id":"https://openalex.org/S4306400115","display_name":"eScholarship (California Digital Library)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I2801248553","host_organization_name":"California Digital Library","host_organization_lineage":["https://openalex.org/I2801248553"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"IEEE Transactions on Parallel and Distributed Systems, vol 36, iss 2","raw_type":"article"}],"best_oa_location":{"id":"pmh:oai:escholarship.org:ark:/13030/qt35n7r10r","is_oa":true,"landing_page_url":"https://escholarship.org/uc/item/35n7r10r","pdf_url":null,"source":{"id":"https://openalex.org/S4306400115","display_name":"eScholarship (California Digital Library)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I2801248553","host_organization_name":"California Digital Library","host_organization_lineage":["https://openalex.org/I2801248553"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"IEEE Transactions on Parallel and Distributed Systems, vol 36, iss 2","raw_type":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":32,"referenced_works":["https://openalex.org/W2018162709","https://openalex.org/W2585029434","https://openalex.org/W2626874786","https://openalex.org/W2896457183","https://openalex.org/W2903557836","https://openalex.org/W2926767350","https://openalex.org/W2946408574","https://openalex.org/W2979826702","https://openalex.org/W3012121966","https://openalex.org/W3086105743","https://openalex.org/W3131944805","https://openalex.org/W3132858873","https://openalex.org/W3158783012","https://openalex.org/W3197525459","https://openalex.org/W3197720002","https://openalex.org/W3210256422","https://openalex.org/W4214658871","https://openalex.org/W4290928031","https://openalex.org/W4367147621","https://openalex.org/W6744307745","https://openalex.org/W6763701032","https://openalex.org/W6763737044","https://openalex.org/W6778883912","https://openalex.org/W6780805062","https://openalex.org/W6788597507","https://openalex.org/W6798473733","https://openalex.org/W6810081322","https://openalex.org/W6810296985","https://openalex.org/W6811340617","https://openalex.org/W6845821369","https://openalex.org/W6850625674","https://openalex.org/W6852481455"],"related_works":["https://openalex.org/W2505380084","https://openalex.org/W4400333498","https://openalex.org/W2086739451","https://openalex.org/W3183233360","https://openalex.org/W1980160788","https://openalex.org/W1656096860","https://openalex.org/W2095928260","https://openalex.org/W2268149564","https://openalex.org/W1984739956","https://openalex.org/W2763312740"],"abstract_inverted_index":{"Characterizing":[0],"and":[1,17,23,33,50,55,63,70,104,121],"predicting":[2,149],"the":[3,29,46,57,64,84,150,209,221],"training":[4,88,152],"performance":[5,107,115,132],"of":[6,48,66,93,126,154,168,182,197,214,231],"modern":[7],"machine":[8],"learning":[9,156],"(ML)":[10],"workloads":[11,142,205],"on":[12,91,144,170,208],"compute":[13,16,80],"systems":[14],"with":[15,83,136,161,164,193],"communication":[18,68,127],"spread":[19],"between":[20,53],"CPUs,":[21],"GPUs,":[22,56],"network":[24,75],"devices":[25,69],"is":[26,212],"not":[27],"only":[28],"key":[30],"to":[31,40,179],"optimization":[32],"planning":[34],"but":[35],"also":[36],"a":[37,165,194,228],"complex":[38],"goal":[39],"achieve.":[41],"The":[42],"primary":[43],"challenges":[44,103],"include":[45],"complexity":[47],"synchronization":[49,139],"load":[51],"balancing":[52],"CPUs":[54],"variance":[58],"in":[59],"input":[60],"data":[61,123],"distribution,":[62],"use":[65],"different":[67],"topologies":[71],"(e.g.,":[72],"NVLink,":[73],"PCIe,":[74],"cards)":[76],"that":[77],"connect":[78],"multiple":[79],"devices,":[81],"coupled":[82],"desire":[85],"for":[86,97,117,140],"flexible":[87],"configurations.":[89],"Built":[90],"top":[92],"our":[94,130,174],"prior":[95],"work":[96],"single-GPU":[98],"platforms,":[99,173],"we":[100],"address":[101],"these":[102],"enable":[105],"multi-GPU":[106,145,172],"modeling<sup":[108],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[109],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">1</sup>":[110],"by":[111],"incorporating":[112],"(1)":[113],"data-distribution-aware":[114],"models":[116,158,160,192],"embedding":[118,223],"table":[119,224],"lookup,":[120],"(2)":[122],"movement":[124],"prediction":[125,175],"collectives,":[128],"into":[129],"upgraded":[131],"modeling":[133],"pipeline":[134,176],"equipped":[135],"inter-and":[137],"intra-rank":[138],"ML":[141,183,204],"trained":[143],"platforms.":[146],"Beyond":[147],"accurately":[148],"per-iteration":[151],"time":[153],"deep":[155],"recommendation":[157],"(DLRM)":[159],"random":[162],"configurations":[163],"geomean":[166,195],"error":[167,196],"5.21%":[169],"two":[171],"generalizes":[177],"well":[178],"other":[180],"types":[181],"workloads,":[184],"such":[185,217],"as":[186,218],"Transformer-based":[187],"natural":[188],"language":[189],"processing":[190],"(NLP)":[191],"3.00%.":[198],"Moreover,":[199],"even":[200],"without":[201],"actually":[202],"running":[203],"like":[206],"DLRMs":[207],"hardware,":[210],"it":[211],"capable":[213],"generating":[215],"insights":[216],"quickly":[219],"selecting":[220],"fastest":[222],"sharding":[225],"configuration":[226],"(with":[227],"success":[229],"rate":[230],"85%).":[232]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":2}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}
