{"id":"https://openalex.org/W4416927400","doi":"https://doi.org/10.1145/3771574","title":"Optimal Scheduling Algorithms for LLM Inference: Theory and Practice","display_name":"Optimal Scheduling Algorithms for LLM Inference: Theory and Practice","publication_year":2025,"publication_date":"2025-12-01","ids":{"openalex":"https://openalex.org/W4416927400","doi":"https://doi.org/10.1145/3771574"},"language":"en","primary_location":{"id":"doi:10.1145/3771574","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3771574","pdf_url":null,"source":{"id":"https://openalex.org/S4210193547","display_name":"Proceedings of the ACM on Measurement and Analysis of Computing Systems","issn_l":"2476-1249","issn":["2476-1249"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"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 ACM on Measurement and Analysis of Computing Systems","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.1145/3771574","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5075448142","display_name":"Agrim Bari","orcid":"https://orcid.org/0009-0003-1550-8683"},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Agrim Bari","raw_affiliation_strings":["Chandra Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA"],"affiliations":[{"raw_affiliation_string":"Chandra Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5036219096","display_name":"Parikshit Hegde","orcid":"https://orcid.org/0000-0003-3477-4947"},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Parikshit Hegde","raw_affiliation_strings":["Chandra Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA"],"affiliations":[{"raw_affiliation_string":"Chandra Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5008917819","display_name":"Gustavo de Veciana","orcid":"https://orcid.org/0000-0002-1498-494X"},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Gustavo de Veciana","raw_affiliation_strings":["Chandra Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA"],"affiliations":[{"raw_affiliation_string":"Chandra Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA","institution_ids":["https://openalex.org/I86519309"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5075448142"],"corresponding_institution_ids":["https://openalex.org/I86519309"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.50160418,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"9","issue":"3","first_page":"1","last_page":"43"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T14347","display_name":"Big Data and Digital Economy","score":0.14239999651908875,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T14347","display_name":"Big Data and Digital Economy","score":0.14239999651908875,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.10610000044107437,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10273","display_name":"IoT and Edge/Fog Computing","score":0.07410000264644623,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.6736999750137329},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.6644999980926514},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5343000292778015},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5342000126838684},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.5116000175476074},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.45080000162124634}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8203999996185303},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6736999750137329},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6644999980926514},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5343000292778015},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5342000126838684},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5116000175476074},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.45080000162124634},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.39169999957084656},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3086000084877014},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2851000130176544},{"id":"https://openalex.org/C142417499","wikidata":"https://www.wikidata.org/wiki/Q331716","display_name":"Amortized analysis","level":3,"score":0.28110000491142273},{"id":"https://openalex.org/C37404715","wikidata":"https://www.wikidata.org/wiki/Q380679","display_name":"Dynamic programming","level":2,"score":0.27619999647140503},{"id":"https://openalex.org/C2984822820","wikidata":"https://www.wikidata.org/wiki/Q1123036","display_name":"Processor scheduling","level":3,"score":0.27559998631477356},{"id":"https://openalex.org/C107568181","wikidata":"https://www.wikidata.org/wiki/Q5319000","display_name":"Dynamic priority scheduling","level":3,"score":0.2718999981880188},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.2703999876976013},{"id":"https://openalex.org/C106516650","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm design","level":2,"score":0.26409998536109924},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.25609999895095825}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3771574","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3771574","pdf_url":null,"source":{"id":"https://openalex.org/S4210193547","display_name":"Proceedings of the ACM on Measurement and Analysis of Computing Systems","issn_l":"2476-1249","issn":["2476-1249"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"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 ACM on Measurement and Analysis of Computing Systems","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1145/3771574","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3771574","pdf_url":null,"source":{"id":"https://openalex.org/S4210193547","display_name":"Proceedings of the ACM on Measurement and Analysis of Computing Systems","issn_l":"2476-1249","issn":["2476-1249"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"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 ACM on Measurement and Analysis of Computing Systems","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3521816219","display_name":null,"funder_award_id":"2148224, CNS-2212202","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":5,"referenced_works":["https://openalex.org/W2081320273","https://openalex.org/W3204998121","https://openalex.org/W4401176373","https://openalex.org/W4408903522","https://openalex.org/W4409047782"],"related_works":[],"abstract_inverted_index":{"With":[0],"the":[1,41,115,150,184,195,199,214,221],"growing":[2],"use":[3],"of":[4],"Large":[5],"Language":[6],"Model":[7],"(LLM)-based":[8],"tools":[9],"like":[10],"ChatGPT,":[11],"Perplexity,":[12],"and":[13,45,64,86,99,120,172,219],"Gemini":[14],"across":[15],"industries,":[16],"there":[17],"is":[18,231,245],"a":[19,32,37,46,54,72,80],"rising":[20],"need":[21],"for":[22,59,62,105],"efficient":[23],"LLM":[24,89,152],"inference":[25,90],"systems.":[26,91],"These":[27],"systems":[28],"handle":[29],"requests":[30,140,163,175],"with":[31,141],"unique":[33],"two-phase":[34],"computation":[35],"structure:":[36],"prefill-phase":[38],"that":[39,48,83,122,164,228],"processes":[40],"full":[42],"input":[43],"prompt":[44,179],"decode-phase":[47],"autoregressively":[49],"generates":[50],"tokens":[51],"one":[52],"at":[53],"time.":[55],"This":[56],"structure":[57],"calls":[58],"new":[60],"strategies":[61],"routing":[63,85],"scheduling":[65,87],"requests.":[66],"In":[67],"this":[68,76],"paper,":[69],"we":[70,113,148],"take":[71],"comprehensive":[73],"approach":[74],"to":[75,160,167,181,210],"challenge":[77],"by":[78,110,217,225],"developing":[79],"theoretical":[81],"framework":[82],"models":[84],"in":[88],"We":[92,191],"identify":[93],"two":[94],"key":[95],"design":[96,149],"principles-optimal":[97],"tiling":[98],"dynamic":[100],"resource":[101],"allocation-that":[102],"are":[103,165],"essential":[104],"achieving":[106],"high":[107],"throughput.":[108],"Guided":[109],"these":[111],"principles,":[112],"propose":[114],"Resource-Aware":[116],"Dynamic":[117],"(RAD)":[118],"scheduler":[119],"prove":[121],"it":[123],"achieves":[124],"throughput":[125],"optimality":[126],"under":[127],"mild":[128],"conditions.":[129],"To":[130,186],"address":[131],"practical":[132],"Service":[133],"Level":[134],"Objectives":[135],"(SLOs)":[136],"such":[137,227],"as":[138],"serving":[139,223],"different":[142],"Time":[143,185],"Between":[144],"Token":[145,188],"(TBT)":[146],"constraints,":[147],"SLO-Aware":[151],"Inference":[153],"(SLAI)":[154],"scheduler.":[155],"SLAI":[156,193,212],"uses":[157],"real-time":[158],"measurements":[159],"prioritize":[161],"decode":[162],"close":[166],"missing":[168],"their":[169],"TBT":[170,238],"deadlines":[171],"reorders":[173],"prefill":[174],"based":[176],"on":[177,194,202],"known":[178],"lengths":[180],"further":[182],"reduce":[183],"First":[187],"(TTFT)":[189],"delays.":[190],"evaluate":[192],"openchat_shareGPT4":[196],"dataset":[197],"using":[198],"Mistral-7B":[200],"model":[201],"an":[203],"NVIDIA":[204],"RTX":[205],"ADA":[206],"6000":[207],"GPU.":[208],"Compared":[209],"Sarathi-Serve,":[211],"reduces":[213],"median":[215,229],"TTFT":[216,230],"53%":[218],"increases":[220],"maximum":[222],"capacity":[224],"26%":[226],"below":[232],"0.5":[233],"seconds,":[234],"while":[235],"meeting":[236],"tail":[237],"latency":[239],"constraints.":[240],"The":[241],"complete":[242],"source":[243],"code":[244],"available":[246],"at:":[247],"https://github.com/agrimUT/SLAI.":[248]},"counts_by_year":[],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-12-02T00:00:00"}