{"id":"https://openalex.org/W4407218342","doi":"https://doi.org/10.1145/3669940.3707215","title":"Helix: Serving Large Language Models over Heterogeneous GPUs and Network via Max-Flow","display_name":"Helix: Serving Large Language Models over Heterogeneous GPUs and Network via Max-Flow","publication_year":2025,"publication_date":"2025-02-06","ids":{"openalex":"https://openalex.org/W4407218342","doi":"https://doi.org/10.1145/3669940.3707215"},"language":"en","primary_location":{"id":"doi:10.1145/3669940.3707215","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3669940.3707215","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 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3669940.3707215","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5051800592","display_name":"Yixuan Mei","orcid":"https://orcid.org/0009-0003-5781-9164"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Yixuan Mei","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA"],"raw_orcid":"https://orcid.org/0009-0003-5781-9164","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076407338","display_name":"Yonghao Zhuang","orcid":null},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yonghao Zhuang","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA"],"raw_orcid":"https://orcid.org/0009-0001-8969-7478","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5015552951","display_name":"Xupeng Miao","orcid":"https://orcid.org/0000-0002-9371-8358"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xupeng Miao","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA"],"raw_orcid":"https://orcid.org/0000-0002-9371-8358","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034227509","display_name":"Juncheng Yang","orcid":"https://orcid.org/0000-0002-0412-1139"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Juncheng Yang","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA"],"raw_orcid":"https://orcid.org/0000-0002-0412-1139","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5062437461","display_name":"Zhihao Jia","orcid":"https://orcid.org/0000-0002-1270-5185"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhihao Jia","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA"],"raw_orcid":"https://orcid.org/0000-0002-1270-5185","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5009725985","display_name":"Rashmi Vinayak","orcid":"https://orcid.org/0000-0002-2227-7460"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rashmi Vinayak","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA"],"raw_orcid":"https://orcid.org/0000-0002-2227-7460","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5051800592"],"corresponding_institution_ids":["https://openalex.org/I74973139"],"apc_list":null,"apc_paid":null,"fwci":14.7,"has_fulltext":false,"cited_by_count":11,"citation_normalized_percentile":{"value":0.98806765,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":96,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"586","last_page":"602"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.9883999824523926,"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"}},"topics":[{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.9883999824523926,"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"}},{"id":"https://openalex.org/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.9767000079154968,"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/T10028","display_name":"Topic Modeling","score":0.9761000275611877,"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":0.7208421230316162},{"id":"https://openalex.org/keywords/flow","display_name":"Flow (mathematics)","score":0.4534272849559784},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.39502546191215515},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.33432239294052124},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.09654828906059265},{"id":"https://openalex.org/keywords/geometry","display_name":"Geometry","score":0.068348228931427}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7208421230316162},{"id":"https://openalex.org/C38349280","wikidata":"https://www.wikidata.org/wiki/Q1434290","display_name":"Flow (mathematics)","level":2,"score":0.4534272849559784},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.39502546191215515},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.33432239294052124},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.09654828906059265},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.068348228931427}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3669940.3707215","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3669940.3707215","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 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3669940.3707215","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3669940.3707215","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 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G5945274387","display_name":null,"funder_award_id":"CNS-2147909,CNS-2211882,CNS-2239351","funder_id":"https://openalex.org/F4320323817","funder_display_name":"Universitas Brawijaya"}],"funders":[{"id":"https://openalex.org/F4320323817","display_name":"Universitas Brawijaya","ror":"https://ror.org/01wk3d929"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":37,"referenced_works":["https://openalex.org/W95608104","https://openalex.org/W2029266474","https://openalex.org/W2114719509","https://openalex.org/W2149294210","https://openalex.org/W2604783387","https://openalex.org/W2734941459","https://openalex.org/W2901299405","https://openalex.org/W2956461999","https://openalex.org/W2981449041","https://openalex.org/W2982157693","https://openalex.org/W3012189484","https://openalex.org/W3013882546","https://openalex.org/W3031276512","https://openalex.org/W3032387272","https://openalex.org/W3033527233","https://openalex.org/W3093783890","https://openalex.org/W4225004481","https://openalex.org/W4281728635","https://openalex.org/W4287116725","https://openalex.org/W4287594180","https://openalex.org/W4308760184","https://openalex.org/W4310561894","https://openalex.org/W4318621310","https://openalex.org/W4321853806","https://openalex.org/W4387321091","https://openalex.org/W4389761026","https://openalex.org/W4391045969","https://openalex.org/W4394945032","https://openalex.org/W4394946008","https://openalex.org/W4395112660","https://openalex.org/W4399911489","https://openalex.org/W4401211704","https://openalex.org/W6778883912","https://openalex.org/W6797086287","https://openalex.org/W6846659131","https://openalex.org/W7046032890","https://openalex.org/W7075998075"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"This":[0,82],"paper":[1],"introduces":[2],"Helix,":[3],"a":[4,38,64],"distributed":[5],"system":[6],"for":[7],"high-throughput,":[8],"low-latency":[9],"large":[10],"language":[11],"model":[12,89],"(LLM)":[13],"serving":[14,119],"in":[15,98],"heterogeneous":[16,32,80,99,106],"GPU":[17,48,54,113],"clusters.":[18],"The":[19],"key":[20],"idea":[21],"behind":[22],"Helix":[23,61,85,117,142],"is":[24,143],"to":[25,71,76,86,111,123,133,139],"formulate":[26],"inference":[27],"computation":[28],"of":[29],"LLMs":[30,78],"over":[31],"GPUs":[33],"and":[34,50,55,91,125,128,135],"network":[35,56],"connections":[36],"as":[37],"max-flow":[39],"problem":[40],"on":[41,79,104],"directed,":[42],"weighted":[43],"graphs,":[44],"whose":[45],"nodes":[46,114],"represent":[47],"instances":[49],"edges":[51],"capture":[52],"both":[53],"heterogeneity":[57],"through":[58],"their":[59],"capacities.":[60],"then":[62],"uses":[63],"mixed":[65],"integer":[66],"linear":[67],"programming":[68],"(MILP)":[69],"algorithm":[70],"discover":[72],"highly":[73,95],"optimized":[74],"strategies":[75],"serve":[77],"GPUs.":[81],"approach":[83],"allows":[84],"jointly":[87],"optimize":[88],"placement":[90],"request":[92],"scheduling,":[93],"two":[94],"entangled":[96],"tasks":[97],"LLM":[100],"serving.":[101],"Our":[102],"evaluation":[103],"several":[105],"clusters":[107],"ranging":[108],"from":[109],"24":[110],"42":[112],"shows":[115],"that":[116],"improves":[118],"throughput":[120],"by":[121,131],"up":[122,132],"3.3x":[124],"reduces":[126],"prompting":[127],"decoding":[129],"latency":[130],"66%":[134],"24%,":[136],"respectively,":[137],"compared":[138],"existing":[140],"approaches.":[141],"available":[144],"at":[145],"https://github.com/Thesys-lab/Helix-ASPLOS25.":[146]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":10}],"updated_date":"2026-04-28T14:05:53.105641","created_date":"2025-10-10T00:00:00"}