{"id":"https://openalex.org/W7124179532","doi":"https://doi.org/10.1109/icpads67057.2025.11322882","title":"HAP: Hybrid Adaptive Parallelism for Efficient Mixture-of-Experts Inference","display_name":"HAP: Hybrid Adaptive Parallelism for Efficient Mixture-of-Experts Inference","publication_year":2025,"publication_date":"2025-12-14","ids":{"openalex":"https://openalex.org/W7124179532","doi":"https://doi.org/10.1109/icpads67057.2025.11322882"},"language":null,"primary_location":{"id":"doi:10.1109/icpads67057.2025.11322882","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icpads67057.2025.11322882","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 31th International Conference on Parallel and Distributed Systems (ICPADS)","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/A5123031524","display_name":"Haoran Lin","orcid":null},"institutions":[{"id":"https://openalex.org/I154099455","display_name":"Shandong University","ror":"https://ror.org/0207yh398","country_code":"CN","type":"education","lineage":["https://openalex.org/I154099455"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Haoran Lin","raw_affiliation_strings":["School of Software, Shandong University,Jinan,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Software, Shandong University,Jinan,China","institution_ids":["https://openalex.org/I154099455"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5021890706","display_name":"Xianzhi Yu","orcid":"https://orcid.org/0000-0002-1497-5525"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xianzhi Yu","raw_affiliation_strings":["Huawei Noah&#x0027;s Ark Lab,Beijing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah&#x0027;s Ark Lab,Beijing,China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103326213","display_name":"Kang Zhao","orcid":"https://orcid.org/0009-0000-3053-3965"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kang Zhao","raw_affiliation_strings":["Huawei Noah&#x0027;s Ark Lab,Beijing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah&#x0027;s Ark Lab,Beijing,China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123039713","display_name":"Han Bao","orcid":null},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Han Bao","raw_affiliation_strings":["Huawei Noah&#x0027;s Ark Lab,Beijing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah&#x0027;s Ark Lab,Beijing,China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Zongyuan Zhan","orcid":null},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zongyuan Zhan","raw_affiliation_strings":["Huawei Noah&#x0027;s Ark Lab,Beijing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah&#x0027;s Ark Lab,Beijing,China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101962045","display_name":"Ting Hu","orcid":"https://orcid.org/0000-0001-7321-0840"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ting Hu","raw_affiliation_strings":["Huawei Noah&#x0027;s Ark Lab,Beijing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah&#x0027;s Ark Lab,Beijing,China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123021625","display_name":"Wulong Liu","orcid":null},"institutions":[{"id":"https://openalex.org/I154099455","display_name":"Shandong University","ror":"https://ror.org/0207yh398","country_code":"CN","type":"education","lineage":["https://openalex.org/I154099455"]},{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wulong Liu","raw_affiliation_strings":["Huawei Noah&#x0027;s Ark Lab,Beijing,China","School of Software, Shandong University,Jinan,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah&#x0027;s Ark Lab,Beijing,China","institution_ids":["https://openalex.org/I2250955327"]},{"raw_affiliation_string":"School of Software, Shandong University,Jinan,China","institution_ids":["https://openalex.org/I154099455"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123038761","display_name":"Zekun Yin","orcid":null},"institutions":[{"id":"https://openalex.org/I154099455","display_name":"Shandong University","ror":"https://ror.org/0207yh398","country_code":"CN","type":"education","lineage":["https://openalex.org/I154099455"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zekun Yin","raw_affiliation_strings":["School of Software, Shandong University,Jinan,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Software, Shandong University,Jinan,China","institution_ids":["https://openalex.org/I154099455"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123045154","display_name":"Xin Li","orcid":null},"institutions":[{"id":"https://openalex.org/I154099455","display_name":"Shandong University","ror":"https://ror.org/0207yh398","country_code":"CN","type":"education","lineage":["https://openalex.org/I154099455"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xin Li","raw_affiliation_strings":["School of Software, Shandong University,Jinan,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Software, Shandong University,Jinan,China","institution_ids":["https://openalex.org/I154099455"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5123017042","display_name":"Weiguo Liu","orcid":null},"institutions":[{"id":"https://openalex.org/I154099455","display_name":"Shandong University","ror":"https://ror.org/0207yh398","country_code":"CN","type":"education","lineage":["https://openalex.org/I154099455"]},{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Weiguo Liu","raw_affiliation_strings":["Huawei Noah&#x0027;s Ark Lab,Beijing,China","School of Software, Shandong University,Jinan,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah&#x0027;s Ark Lab,Beijing,China","institution_ids":["https://openalex.org/I2250955327"]},{"raw_affiliation_string":"School of Software, Shandong University,Jinan,China","institution_ids":["https://openalex.org/I154099455"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":10,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.7588,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.91384931,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"8"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12814","display_name":"Gaussian Processes and Bayesian Inference","score":0.36070001125335693,"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/T12814","display_name":"Gaussian Processes and Bayesian Inference","score":0.36070001125335693,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.2093999981880188,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.10819999873638153,"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/inference","display_name":"Inference","score":0.8328999876976013},{"id":"https://openalex.org/keywords/flexibility","display_name":"Flexibility (engineering)","score":0.5551000237464905},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.5188000202178955},{"id":"https://openalex.org/keywords/decomposition","display_name":"Decomposition","score":0.4043000042438507},{"id":"https://openalex.org/keywords/parallelism","display_name":"Parallelism (grammar)","score":0.37689998745918274},{"id":"https://openalex.org/keywords/computational-model","display_name":"Computational model","score":0.337799996137619},{"id":"https://openalex.org/keywords/computational-complexity-theory","display_name":"Computational complexity theory","score":0.2912999987602234}],"concepts":[{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.8328999876976013},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7793999910354614},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.5551000237464905},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.5188000202178955},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4442000091075897},{"id":"https://openalex.org/C124681953","wikidata":"https://www.wikidata.org/wiki/Q339062","display_name":"Decomposition","level":2,"score":0.4043000042438507},{"id":"https://openalex.org/C2781172179","wikidata":"https://www.wikidata.org/wiki/Q853109","display_name":"Parallelism (grammar)","level":2,"score":0.37689998745918274},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.337799996137619},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.32580000162124634},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3246000111103058},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3102000057697296},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.2912999987602234},{"id":"https://openalex.org/C56086750","wikidata":"https://www.wikidata.org/wiki/Q6042592","display_name":"Integer programming","level":2,"score":0.2904999852180481},{"id":"https://openalex.org/C157170001","wikidata":"https://www.wikidata.org/wiki/Q4781507","display_name":"Applications of artificial intelligence","level":2,"score":0.28769999742507935},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.2849000096321106},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.28060001134872437},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.2732999920845032},{"id":"https://openalex.org/C143724316","wikidata":"https://www.wikidata.org/wiki/Q312468","display_name":"Series (stratigraphy)","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.26820001006126404},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.25459998846054077},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.25099998712539673}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icpads67057.2025.11322882","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icpads67057.2025.11322882","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 31th International Conference on Parallel and Distributed Systems (ICPADS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.5422478318214417,"display_name":"Industry, innovation and infrastructure","id":"https://metadata.un.org/sdg/9"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":8,"referenced_works":["https://openalex.org/W3093933627","https://openalex.org/W3094271339","https://openalex.org/W4289828103","https://openalex.org/W4376652719","https://openalex.org/W4385567093","https://openalex.org/W4387321091","https://openalex.org/W4396817294","https://openalex.org/W4415797413"],"related_works":[],"abstract_inverted_index":{"Current":[0],"inference":[1,23,57,87,122,150,158],"systems":[2],"for":[3,101],"Mixture-of-Experts":[4],"(MoE)":[5],"models":[6],"primarily":[7],"employ":[8],"static":[9,14],"parallelization":[10],"strategies.":[11,105],"However,":[12],"these":[13],"approaches":[15],"cannot":[16],"consistently":[17,133],"achieve":[18,138],"optimal":[19,116],"performance":[20,142,188],"across":[21,190],"different":[22],"scenarios,":[24],"as":[25],"they":[26],"lack":[27],"the":[28,75,79,94,115,144,154],"flexibility":[29],"to":[30,32,54,120,143,153],"adapt":[31],"varying":[33,125],"computational":[34,73,126],"requirements.":[35],"In":[36],"this":[37],"work,":[38],"we":[39],"propose":[40],"HAP":[41,63,112,132,182],"(Hybrid":[42],"Adaptive":[43],"Parallelism),":[44],"a":[45,85,97],"novel":[46],"method":[47],"that":[48,131,137],"dynamically":[49],"selects":[50],"hybrid":[51,117],"parallel":[52,104,118,135],"strategies":[53],"enhance":[55],"MoE":[56,68,192],"efficiency.":[58],"The":[59],"fundamental":[60],"innovation":[61],"of":[62,96,161],"lies":[64],"in":[65,148],"hierarchically":[66],"decomposing":[67],"architectures":[69],"into":[70],"two":[71],"distinct":[72],"modules:":[74],"Attention":[76],"module":[77],"and":[78,168,176,197],"Expert":[80],"module,":[81],"each":[82],"augmented":[83],"with":[84],"specialized":[86],"latency":[88],"simulation":[89],"model.":[90],"This":[91],"decomposition":[92],"promotes":[93],"construction":[95],"comprehensive":[98],"search":[99],"space":[100],"seeking":[102],"model":[103,193],"By":[106],"leveraging":[107],"Integer":[108],"Linear":[109],"Programming":[110],"(ILP),":[111],"could":[113],"solve":[114],"configurations":[119,136],"maximize":[121],"efficiency":[123],"under":[124],"constraints.":[127],"Our":[128],"experiments":[129],"demonstrate":[130],"determines":[134],"comparable":[139],"or":[140],"superior":[141],"TP":[145],"strategy":[146],"prevalent":[147],"mainstream":[149],"systems.":[151],"Compared":[152],"TP-based":[155],"inference,":[156],"HAP-based":[157],"achieves":[159],"speedups":[160],"<tex":[162,169],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[163,170],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$1.68":[164],"\\times,":[165],"1.77":[166],"\\times$</tex>,":[167],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$1.57":[171],"\\times$</tex>":[172],"on":[173],"A100,":[174],"A6000,":[175],"V100":[177],"GPU":[178],"platforms,":[179],"respectively.":[180],"Furthermore,":[181],"showcases":[183],"remarkable":[184],"generalization":[185],"capability,":[186],"maintaining":[187],"effectiveness":[189],"diverse":[191],"configurations,":[194],"including":[195],"Mixtral":[196],"Qwen":[198],"series":[199],"models.":[200]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-01-15T00:00:00"}