{"id":"https://openalex.org/W4399452094","doi":"https://doi.org/10.1145/3662158.3662806","title":"System Optimizations for Enabling Training of Extreme Long Sequence Transformer Models","display_name":"System Optimizations for Enabling Training of Extreme Long Sequence Transformer Models","publication_year":2024,"publication_date":"2024-06-05","ids":{"openalex":"https://openalex.org/W4399452094","doi":"https://doi.org/10.1145/3662158.3662806"},"language":"en","primary_location":{"id":"doi:10.1145/3662158.3662806","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3662158.3662806","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 43rd ACM Symposium on Principles of Distributed Computing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3662158.3662806","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5017294111","display_name":"Sam Ad\u00e9 Jacobs","orcid":"https://orcid.org/0000-0003-3425-5602"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Sam Ade Jacobs","raw_affiliation_strings":["Microsoft Inc, Redmond, United States of America"],"affiliations":[{"raw_affiliation_string":"Microsoft Inc, Redmond, United States of America","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103228553","display_name":"Masahiro Tanaka","orcid":"https://orcid.org/0009-0001-3972-8133"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Masahiro Tanaka","raw_affiliation_strings":["Microsoft Inc, Redmond, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Inc, Redmond, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100691056","display_name":"Chengming Zhang","orcid":"https://orcid.org/0000-0003-3008-9133"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chengming Zhang","raw_affiliation_strings":["Microsoft Inc, Redmond, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Inc, Redmond, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077768924","display_name":"Minjia Zhang","orcid":"https://orcid.org/0000-0002-8165-166X"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Minjia Zhang","raw_affiliation_strings":["Microsoft Inc, Redmond, United States of America"],"affiliations":[{"raw_affiliation_string":"Microsoft Inc, Redmond, United States of America","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108940878","display_name":"Reza Yazdani Aminadabi","orcid":null},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Reza Yazdani Aminadabi","raw_affiliation_strings":["Microsoft Inc, Redmond, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Inc, Redmond, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043209884","display_name":"Shuaiwen Leon Song","orcid":"https://orcid.org/0000-0002-8402-1436"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Shuaiwen Leon Song","raw_affiliation_strings":["Microsoft Inc, Redmond, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Inc, Redmond, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069595628","display_name":"Samyam Rajbhandari","orcid":"https://orcid.org/0000-0002-0386-8759"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Samyam Rajbhandari","raw_affiliation_strings":["Microsoft Inc, Redmond, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Inc, Redmond, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5040302174","display_name":"Yuxiong He","orcid":"https://orcid.org/0000-0003-0478-8854"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yuxiong He","raw_affiliation_strings":["Microsoft Inc, Redmond, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Inc, Redmond, USA","institution_ids":["https://openalex.org/I1290206253"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5017294111"],"corresponding_institution_ids":["https://openalex.org/I1290206253"],"apc_list":null,"apc_paid":null,"fwci":3.6261,"has_fulltext":false,"cited_by_count":10,"citation_normalized_percentile":{"value":0.93607546,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"121","last_page":"130"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T13650","display_name":"Computational Physics and Python Applications","score":0.9800000190734863,"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"}},"topics":[{"id":"https://openalex.org/T13650","display_name":"Computational Physics and Python Applications","score":0.9800000190734863,"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/T10271","display_name":"Seismic Imaging and Inversion Techniques","score":0.9790999889373779,"subfield":{"id":"https://openalex.org/subfields/1908","display_name":"Geophysics"},"field":{"id":"https://openalex.org/fields/19","display_name":"Earth and Planetary Sciences"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11801","display_name":"Reservoir Engineering and Simulation Methods","score":0.9768999814987183,"subfield":{"id":"https://openalex.org/subfields/2212","display_name":"Ocean 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":0.7146562337875366},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.6118601560592651},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.49446606636047363},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.47937873005867004},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3967253565788269},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.16020548343658447},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.11190539598464966},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.1054181456565857}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7146562337875366},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6118601560592651},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.49446606636047363},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.47937873005867004},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3967253565788269},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.16020548343658447},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.11190539598464966},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.1054181456565857},{"id":"https://openalex.org/C54355233","wikidata":"https://www.wikidata.org/wiki/Q7162","display_name":"Genetics","level":1,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"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":1,"locations":[{"id":"doi:10.1145/3662158.3662806","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3662158.3662806","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 43rd ACM Symposium on Principles of Distributed Computing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3662158.3662806","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3662158.3662806","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 43rd ACM Symposium on Principles of Distributed Computing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":7,"referenced_works":["https://openalex.org/W2747329762","https://openalex.org/W2969388332","https://openalex.org/W2972087877","https://openalex.org/W3129831491","https://openalex.org/W3132259035","https://openalex.org/W3163073193","https://openalex.org/W3205803342"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2748952813","https://openalex.org/W230091440","https://openalex.org/W2390279801","https://openalex.org/W2233261550","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2810751659","https://openalex.org/W258997015","https://openalex.org/W2376932109"],"abstract_inverted_index":{"Computation":[0],"in":[1,91],"a":[2,113],"typical":[3],"Transformer-based":[4],"large":[5,105],"language":[6],"model":[7,52],"(LLM)":[8],"can":[9],"be":[10],"characterized":[11],"by":[12,96],"batch":[13,41],"size,":[14,42,47],"hidden":[15,46],"dimension,":[16],"number":[17],"of":[18,60],"layers,":[19],"and":[20,48,116,123,143,176,196],"sequence":[21,69,78,86,92,104,130,141,165,174,203],"length.":[22,131],"Until":[23],"now,":[24],"system":[25],"works":[26,90],"for":[27,40,45,51,67,76,119,150],"accelerating":[28],"LLM":[29,125],"training":[30,126],"have":[31],"focused":[32],"on":[33],"the":[34,140,206],"first":[35],"three":[36],"dimensions:":[37],"data":[38,138],"parallelism":[39,44,50,61,93],"tensor":[43],"pipeline":[49],"depth":[53],"or":[54,65],"layers.":[55],"These":[56],"widely":[57],"studied":[58],"forms":[59],"are":[62,82,94,179],"not":[63],"targeted":[64],"optimized":[66],"long":[68,77,103,129],"Transformer":[70],"models.":[71,106],"Given":[72],"practical":[73],"application":[74],"needs":[75],"LLM,":[79],"renewed":[80],"attentions":[81],"being":[83],"drawn":[84],"to":[85,102,189],"parallelism.":[87],"However,":[88],"existing":[89,207],"constrained":[95],"memory-communication":[97],"inefficiency,":[98],"limiting":[99],"their":[100],"scalability":[101],"In":[107],"this":[108],"work,":[109],"we":[110],"introduce":[111],"Ulysses,":[112],"novel,":[114],"portable,":[115],"effective":[117],"methodology":[118],"enabling":[120],"highly":[121],"efficient":[122,146],"scalable":[124],"with":[127,200],"extremely":[128],"Ulysses":[132,168,187],"at":[133],"its":[134],"core":[135],"partitions":[136],"input":[137],"along":[139],"dimension":[142],"employs":[144],"an":[145],"all-to-all":[147],"collective":[148],"communication":[149,154,162,171],"attention":[151],"computation.":[152],"Theoretical":[153],"analysis":[155],"shows":[156],"that,":[157],"whereas":[158],"other":[159],"methods":[160],"incur":[161],"overhead":[163],"as":[164],"length":[166,175,195,204],"increases,":[167],"maintains":[169],"constant":[170],"volume":[172],"when":[173],"compute":[177],"devices":[178],"increased":[180],"proportionally.":[181],"Furthermore,":[182],"experimental":[183],"evaluations":[184],"show":[185],"that":[186],"scales":[188],"more":[190],"than":[191,205],"1":[192],"million":[193],"context":[194],"trains":[197],"2.5x":[198],"faster":[199],"4x":[201],"longer":[202],"method":[208],"SOTA":[209],"baseline.":[210]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":1}],"updated_date":"2025-12-26T23:08:49.675405","created_date":"2025-10-10T00:00:00"}