{"id":"https://openalex.org/W4404401017","doi":"https://doi.org/10.1145/3694715.3695948","title":"LoongServe: Efficiently Serving Long-Context Large Language Models with Elastic Sequence Parallelism","display_name":"LoongServe: Efficiently Serving Long-Context Large Language Models with Elastic Sequence Parallelism","publication_year":2024,"publication_date":"2024-11-04","ids":{"openalex":"https://openalex.org/W4404401017","doi":"https://doi.org/10.1145/3694715.3695948"},"language":"en","primary_location":{"id":"doi:10.1145/3694715.3695948","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3694715.3695948","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3694715.3695948","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM SIGOPS 30th Symposium on Operating Systems Principles","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3694715.3695948","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100660218","display_name":"Bingyang Wu","orcid":"https://orcid.org/0000-0001-7221-8007"},"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":"Bingyang Wu","raw_affiliation_strings":["School of Computer Science, Peking University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0001-7221-8007","affiliations":[{"raw_affiliation_string":"School of Computer Science, Peking University, Beijing, China","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5106305638","display_name":"Shengyu Liu","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":"Shengyu Liu","raw_affiliation_strings":["School of Computer Science, Peking University, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0002-9182-4090","affiliations":[{"raw_affiliation_string":"School of Computer Science, Peking University, Beijing, China","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000644890","display_name":"Yinmin Zhong","orcid":"https://orcid.org/0000-0002-2504-7652"},"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":"Yinmin Zhong","raw_affiliation_strings":["School of Computer Science, Peking University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-2504-7652","affiliations":[{"raw_affiliation_string":"School of Computer Science, Peking University, Beijing, China","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100614087","display_name":"Peng Sun","orcid":"https://orcid.org/0000-0001-8456-0491"},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Peng Sun","raw_affiliation_strings":["Shanghai AI Lab, Shanghai, China"],"raw_orcid":"https://orcid.org/0000-0001-8456-0491","affiliations":[{"raw_affiliation_string":"Shanghai AI Lab, Shanghai, China","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5052249316","display_name":"Xuanzhe Liu","orcid":"https://orcid.org/0000-0002-7908-8484"},"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":"Xuanzhe Liu","raw_affiliation_strings":["School of Computer Science, Peking University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-7908-8484","affiliations":[{"raw_affiliation_string":"School of Computer Science, Peking University, Beijing, China","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101882004","display_name":"Xin Jin","orcid":"https://orcid.org/0000-0001-8741-5847"},"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":"Xin Jin","raw_affiliation_strings":["School of Computer Science, Peking University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0001-8741-5847","affiliations":[{"raw_affiliation_string":"School of Computer Science, Peking University, Beijing, China","institution_ids":["https://openalex.org/I20231570"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":9.7246,"has_fulltext":false,"cited_by_count":35,"citation_normalized_percentile":{"value":0.98474759,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"640","last_page":"654"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9997000098228455,"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/T10028","display_name":"Topic Modeling","score":0.9997000098228455,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9983000159263611,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9914000034332275,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8298245072364807},{"id":"https://openalex.org/keywords/parallelism","display_name":"Parallelism (grammar)","score":0.7684763669967651},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6636770367622375},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.6438103318214417},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5531274676322937},{"id":"https://openalex.org/keywords/data-parallelism","display_name":"Data parallelism","score":0.47802814841270447},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.4016479253768921},{"id":"https://openalex.org/keywords/geology","display_name":"Geology","score":0.05552351474761963}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8298245072364807},{"id":"https://openalex.org/C2781172179","wikidata":"https://www.wikidata.org/wiki/Q853109","display_name":"Parallelism (grammar)","level":2,"score":0.7684763669967651},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6636770367622375},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.6438103318214417},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5531274676322937},{"id":"https://openalex.org/C61483411","wikidata":"https://www.wikidata.org/wiki/Q3124522","display_name":"Data parallelism","level":3,"score":0.47802814841270447},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.4016479253768921},{"id":"https://openalex.org/C127313418","wikidata":"https://www.wikidata.org/wiki/Q1069","display_name":"Geology","level":0,"score":0.05552351474761963},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","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/C54355233","wikidata":"https://www.wikidata.org/wiki/Q7162","display_name":"Genetics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3694715.3695948","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3694715.3695948","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3694715.3695948","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM SIGOPS 30th Symposium on Operating Systems Principles","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3694715.3695948","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3694715.3695948","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3694715.3695948","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM SIGOPS 30th Symposium on Operating Systems Principles","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G4497957390","display_name":null,"funder_award_id":"62172008","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G7732961354","display_name":null,"funder_award_id":"2022YFB4500700","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"},{"id":"https://openalex.org/G8414340155","display_name":null,"funder_award_id":"62325201","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"},{"id":"https://openalex.org/F4320335777","display_name":"National Key Research and Development Program of China","ror":null}],"has_content":{"pdf":true,"grobid_xml":false},"content_urls":{"pdf":"https://content.openalex.org/works/W4404401017.pdf"},"referenced_works_count":6,"referenced_works":["https://openalex.org/W2009319645","https://openalex.org/W2798515322","https://openalex.org/W2954698171","https://openalex.org/W4385571616","https://openalex.org/W4389518760","https://openalex.org/W4401018280"],"related_works":["https://openalex.org/W2950520577","https://openalex.org/W1501159154","https://openalex.org/W1554644772","https://openalex.org/W2003935582","https://openalex.org/W2494130044","https://openalex.org/W3170887803","https://openalex.org/W74409296","https://openalex.org/W3209384898","https://openalex.org/W4230999561","https://openalex.org/W1595672120"],"abstract_inverted_index":{"The":[0],"context":[1],"window":[2],"of":[3,28,101],"large":[4],"language":[5],"models":[6],"(LLMs)":[7],"is":[8],"rapidly":[9],"increasing,":[10],"leading":[11],"to":[12,47,68,71,149,152,158],"a":[13,60],"huge":[14],"variance":[15,73],"in":[16,51,103],"resource":[17],"usage":[18],"between":[19,25],"different":[20,26,52,75],"requests":[21,50,76],"as":[22,24],"well":[23],"phases":[27],"the":[29,44,72,99,145],"same":[30],"request.":[31],"Restricted":[32],"by":[33,96,109,128,147],"static":[34],"parallelism":[35,62,66,102],"strategies,":[36],"existing":[37],"LLM":[38,88],"serving":[39,89],"systems":[40],"cannot":[41],"efficiently":[42],"utilize":[43],"underlying":[45],"resources":[46],"serve":[48],"variable-length":[49],"phases.":[53,78],"To":[54],"address":[55],"this":[56],"problem,":[57],"we":[58,82],"propose":[59],"new":[61],"paradigm,":[63],"elastic":[64],"sequence":[65],"(ESP),":[67],"elastically":[69,97],"adapt":[70],"across":[74,133],"and":[77,84,115,122,155],"Based":[79],"on":[80],"ESP,":[81],"design":[83],"build":[85],"LoongServe,":[86],"an":[87],"system":[90],"that":[91,142],"(1)":[92],"improves":[93,106,124,144],"computation":[94],"efficiency":[95,108,127],"adjusting":[98],"degree":[100],"real-time,":[104],"(2)":[105],"communication":[107,119],"reducing":[110,129],"key-value":[111,130],"cache":[112,131],"migration":[113],"overhead":[114],"overlapping":[116],"partial":[117],"decoding":[118],"with":[120],"computation,":[121],"(3)":[123],"GPU":[125],"memory":[126],"fragmentation":[132],"instances.":[134],"Our":[135],"evaluation":[136],"under":[137],"diverse":[138],"real-world":[139],"datasets":[140],"shows":[141],"LoongServe":[143],"throughput":[146],"up":[148],"3.85\u00d7":[150],"compared":[151,157],"chunked":[153],"prefill":[154],"5.81\u00d7":[156],"prefill-decoding":[159],"disaggregation.":[160]},"counts_by_year":[{"year":2026,"cited_by_count":10},{"year":2025,"cited_by_count":23},{"year":2024,"cited_by_count":2}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}