{"id":"https://openalex.org/W4415399006","doi":"https://doi.org/10.1109/tcad.2025.3624135","title":"AsyncGrid: An Intralayer and Interlayer Asynchronous Hybrid Parallelism System for Responsive Edge LLM Inference","display_name":"AsyncGrid: An Intralayer and Interlayer Asynchronous Hybrid Parallelism System for Responsive Edge LLM Inference","publication_year":2025,"publication_date":"2025-10-21","ids":{"openalex":"https://openalex.org/W4415399006","doi":"https://doi.org/10.1109/tcad.2025.3624135"},"language":null,"primary_location":{"id":"doi:10.1109/tcad.2025.3624135","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcad.2025.3624135","pdf_url":null,"source":{"id":"https://openalex.org/S100835903","display_name":"IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems","issn_l":"0278-0070","issn":["0278-0070","1937-4151"],"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 Computer-Aided Design of Integrated Circuits and Systems","raw_type":"journal-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/A5035249009","display_name":"Yi Xiong","orcid":"https://orcid.org/0009-0002-7439-2505"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yi Xiong","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"],"raw_orcid":"https://orcid.org/0009-0002-7439-2505","affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100421958","display_name":"Rui Zhang","orcid":"https://orcid.org/0009-0000-9961-1108"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rui Zhang","raw_affiliation_strings":["Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5117084700","display_name":"Yulong Zu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yulong Zu","raw_affiliation_strings":["Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101644806","display_name":"Weihong Liu","orcid":"https://orcid.org/0000-0002-0308-0259"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Weihong Liu","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"],"raw_orcid":"https://orcid.org/0000-0002-0308-0259","affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022299432","display_name":"Zongwei Zhu","orcid":"https://orcid.org/0000-0003-3607-2631"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zongwei Zhu","raw_affiliation_strings":["Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China"],"raw_orcid":"https://orcid.org/0000-0003-3607-2631","affiliations":[{"raw_affiliation_string":"Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101916933","display_name":"Jiawei Geng","orcid":"https://orcid.org/0000-0001-7405-9820"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiawei Geng","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"],"raw_orcid":"https://orcid.org/0000-0001-7405-9820","affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103281432","display_name":"Boyu Li","orcid":"https://orcid.org/0009-0001-2796-2686"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Boyu Li","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"],"raw_orcid":"https://orcid.org/0009-0001-2796-2686","affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101273644","display_name":"Qianyue Cao","orcid":null},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Qianyue Cao","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"],"raw_orcid":"https://orcid.org/0009-0001-2861-8211","affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5077322091","display_name":"Xuehai Zhou","orcid":"https://orcid.org/0000-0002-8360-3143"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xuehai Zhou","raw_affiliation_strings":["School of Computer Science and Technology, University of Science and Technology of China, Hefei, China"],"raw_orcid":"https://orcid.org/0000-0002-8360-3143","affiliations":[{"raw_affiliation_string":"School of Computer Science and Technology, University of Science and Technology of China, Hefei, China","institution_ids":["https://openalex.org/I126520041"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":9,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.25779098,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"45","issue":"6","first_page":"2826","last_page":"2839"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10232","display_name":"Optical Network Technologies","score":0.9940999746322632,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10232","display_name":"Optical Network Technologies","score":0.9940999746322632,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12146","display_name":"Power Line Communications and Noise","score":0.9860000014305115,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10847","display_name":"Advanced Optical Network Technologies","score":0.9635999798774719,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic 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/asynchronous-communication","display_name":"Asynchronous communication","score":0.7718999981880188},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.57669997215271},{"id":"https://openalex.org/keywords/task-parallelism","display_name":"Task parallelism","score":0.492900013923645},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.48159998655319214},{"id":"https://openalex.org/keywords/concurrency","display_name":"Concurrency","score":0.4778999984264374},{"id":"https://openalex.org/keywords/data-parallelism","display_name":"Data parallelism","score":0.4749999940395355},{"id":"https://openalex.org/keywords/parallelism","display_name":"Parallelism (grammar)","score":0.46720001101493835},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.428600013256073}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.849399983882904},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.7718999981880188},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.57669997215271},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5410000085830688},{"id":"https://openalex.org/C42992933","wikidata":"https://www.wikidata.org/wiki/Q691169","display_name":"Task parallelism","level":3,"score":0.492900013923645},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.48159998655319214},{"id":"https://openalex.org/C193702766","wikidata":"https://www.wikidata.org/wiki/Q1414548","display_name":"Concurrency","level":2,"score":0.4778999984264374},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.47519999742507935},{"id":"https://openalex.org/C61483411","wikidata":"https://www.wikidata.org/wiki/Q3124522","display_name":"Data parallelism","level":3,"score":0.4749999940395355},{"id":"https://openalex.org/C2781172179","wikidata":"https://www.wikidata.org/wiki/Q853109","display_name":"Parallelism (grammar)","level":2,"score":0.46720001101493835},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.428600013256073},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.4140999913215637},{"id":"https://openalex.org/C140763907","wikidata":"https://www.wikidata.org/wiki/Q2714055","display_name":"Instruction-level parallelism","level":3,"score":0.36970001459121704},{"id":"https://openalex.org/C22684755","wikidata":"https://www.wikidata.org/wiki/Q847526","display_name":"Queueing theory","level":2,"score":0.36890000104904175},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.352400004863739},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3346000015735626},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.33219999074935913},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.3264999985694885},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.32330000400543213},{"id":"https://openalex.org/C3543717","wikidata":"https://www.wikidata.org/wiki/Q6007302","display_name":"Implicit parallelism","level":4,"score":0.3181000053882599},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.30149999260902405},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.2996000051498413},{"id":"https://openalex.org/C160403385","wikidata":"https://www.wikidata.org/wiki/Q220543","display_name":"Queue","level":2,"score":0.2858000099658966}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tcad.2025.3624135","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcad.2025.3624135","pdf_url":null,"source":{"id":"https://openalex.org/S100835903","display_name":"IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems","issn_l":"0278-0070","issn":["0278-0070","1937-4151"],"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 Computer-Aided Design of Integrated Circuits and Systems","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G2050173416","display_name":null,"funder_award_id":"2022YFB4501600","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"},{"id":"https://openalex.org/G3723498926","display_name":null,"funder_award_id":"YN2260080008","funder_id":"https://openalex.org/F4320324873","funder_display_name":"South University of Science and Technology of China"}],"funders":[{"id":"https://openalex.org/F4320324873","display_name":"South University of Science and Technology of China","ror":"https://ror.org/049tv2d57"},{"id":"https://openalex.org/F4320335777","display_name":"National Key Research and Development Program of China","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Edge":[0],"deployment":[1],"of":[2,31,40,51,74,155,226,267],"large":[3],"language":[4],"models":[5,246],"(LLMs)":[6],"is":[7,83],"increasingly":[8],"attractive":[9],"due":[10,119,141],"to":[11,95,120,142,173,181,210,260],"its":[12],"advantages":[13,292],"in":[14,25,88,160,190,299],"privacy,":[15],"customization,":[16],"and":[17,35,63,162,213,274,302],"availability.":[18],"However,":[19],"edge":[20,65,96,241,287],"environments":[21,97],"face":[22],"significant":[23,291],"challenges":[24],"reducing":[26],"Time-to-First-Token":[27],"(TTFT).":[28],"TTFT":[29,219],"consists":[30],"(1)":[32,245],"queuing":[33,79],"delay":[34],"(2)":[36,253],"prefill":[37,61,68,107,117,130,139,175,183],"latency,":[38],"both":[39,300],"which":[41,115,168],"are":[42],"exacerbated":[43],"by":[44],"edge\u2011resource":[45],"constraints:":[46],"the":[47,71,153,156,223,265],"substantial":[48,297],"computational":[49],"demands":[50],"LLM":[52,235,242],"inference":[53,236],"grow":[54],"superlinearly":[55],"with":[56,197,264],"prompt":[57],"length,":[58],"causing":[59],"high":[60,138],"latency;":[62],"limited":[64],"resources":[66],"restrict":[67],"throughput,":[69],"preventing":[70],"timely":[72],"handling":[73],"incoming":[75],"requests,":[76],"thereby":[77],"exacerbating":[78],"delays.":[80],"Model":[81],"parallelism":[82,101,125,200,208],"a":[84,202,250],"commonly":[85],"used":[86],"solution":[87],"cloud-based":[89],"systems,":[90],"but":[91,109,136],"directly":[92],"applying":[93],"it":[94],"proves":[98],"ineffective.":[99],"Intra-layer":[100,164],"(e.g.,":[102,126],"tensor/sequence":[103],"parallelism)":[104,128],"can":[105],"reduce":[106,174],"latency":[108,140,176,269],"suffers":[110],"from":[111],"frequent":[112],"global":[113,179],"synchronization,":[114],"bottlenecks":[116],"throughput":[118,131,184,272],"edge-limited":[121],"interconnection":[122],"bandwidth.":[123],"Inter-layer":[124],"pipeline":[127],"improves":[129],"via":[132],"fully":[133],"asynchronous":[134,199],"execution":[135,262],"retains":[137],"stage-wise":[143],"serialized":[144],"computation.":[145],"To":[146,221],"address":[147],"this":[148,150,193,227,230],"dilemma,":[149],"paper":[151,194,231],"leverages":[152],"properties":[154],"causal":[157],"attention":[158],"mechanism":[159],"LLMs":[161],"proposes":[163],"Asynchronous":[165],"Parallelism":[166],"(IAP),":[167],"performs":[169],"intra-layer":[170,191],"parallel":[171],"computations":[172],"while":[177,270],"avoiding":[178],"synchronization":[180],"mitigate":[182],"bottlenecks.":[185],"Moreover,":[186],"considering":[187],"communication":[188,278],"sensitivity":[189],"parallelism,":[192,229],"integrate":[195],"IAP":[196],"inter-layer":[198],"into":[201],"unified":[203],"plan":[204],"space.":[205],"This":[206],"hybrid":[207,228],"adapts":[209],"diverse":[211],"hardware":[212],"request":[214],"loads,":[215],"enabling":[216],"more":[217],"effective":[218],"optimization.":[220],"enable":[222],"end-to-end":[224],"implementation":[225],"propose":[232],"AsyncGrid,":[233],"an":[234,255,286],"system":[237],"tailored":[238],"for":[239],"responsive":[240],"inference.":[243],"AsyncGrid":[244],"runtime":[247],"overheads":[248],"through":[249],"performance":[251],"profiler,":[252],"employs":[254],"integer":[256],"programming":[257],"(IP)":[258],"formulation":[259],"optimize":[261],"plan,":[263],"objective":[266],"minimizing":[268],"meeting":[271],"requirements,":[273],"(3)":[275],"implements":[276],"fine-grained":[277],"optimization":[279],"during":[280],"runtime.":[281],"A":[282],"comprehensive":[283],"evaluation":[284],"on":[285],"testbed":[288],"demonstrates":[289],"AsyncGrid\u2019s":[290],"over":[293],"existing":[294],"methods,":[295],"achieving":[296],"improvements":[298],"homogeneous":[301],"heterogeneous":[303],"settings.":[304]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-22T00:00:00"}
