{"id":"https://openalex.org/W7110859006","doi":"https://doi.org/10.1145/3774899.3775014","title":"Elastic MIG Reconfiguration with PCIe-Aware Placement for Multi-Tenant GPUs","display_name":"Elastic MIG Reconfiguration with PCIe-Aware Placement for Multi-Tenant GPUs","publication_year":2025,"publication_date":"2025-12-09","ids":{"openalex":"https://openalex.org/W7110859006","doi":"https://doi.org/10.1145/3774899.3775014"},"language":null,"primary_location":{"id":"doi:10.1145/3774899.3775014","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3774899.3775014","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 11th International Workshop on Serverless Computing","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":null,"display_name":"Erfan Darzi","orcid":"https://orcid.org/0000-0002-6003-0119"},"institutions":[{"id":"https://openalex.org/I136199984","display_name":"Harvard University","ror":"https://ror.org/03vek6s52","country_code":"US","type":"education","lineage":["https://openalex.org/I136199984"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Erfan Darzi","raw_affiliation_strings":["Harvard University, Boston, MA, USA"],"raw_orcid":"https://orcid.org/0000-0002-6003-0119","affiliations":[{"raw_affiliation_string":"Harvard University, Boston, MA, USA","institution_ids":["https://openalex.org/I136199984"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Sree Bhargavi Balija","orcid":"https://orcid.org/0009-0001-0192-9925"},"institutions":[{"id":"https://openalex.org/I160856358","display_name":"University of San Diego","ror":"https://ror.org/03jbbze48","country_code":"US","type":"education","lineage":["https://openalex.org/I160856358"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sree Bhargavi Balija","raw_affiliation_strings":["University of California, San Diego, San diego, USA"],"raw_orcid":"https://orcid.org/0009-0001-0192-9925","affiliations":[{"raw_affiliation_string":"University of California, San Diego, San diego, USA","institution_ids":["https://openalex.org/I160856358"]}]},{"author_position":"last","author":{"id":null,"display_name":"Buse Demir","orcid":"https://orcid.org/0000-0002-2246-2471"},"institutions":[{"id":"https://openalex.org/I2801851002","display_name":"Harvard University Press","ror":"https://ror.org/006v7bf86","country_code":"US","type":"other","lineage":["https://openalex.org/I136199984","https://openalex.org/I2801851002"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Buse Demir","raw_affiliation_strings":["Harvard University, Boston, USA"],"raw_orcid":"https://orcid.org/0000-0002-2246-2471","affiliations":[{"raw_affiliation_string":"Harvard University, Boston, USA","institution_ids":["https://openalex.org/I2801851002"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I136199984"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.61922879,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"20","last_page":"25"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.4027999937534332,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.4027999937534332,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.1362999975681305,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10829","display_name":"Interconnection Networks and Systems","score":0.06809999793767929,"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/latency","display_name":"Latency (audio)","score":0.7134000062942505},{"id":"https://openalex.org/keywords/cloud-computing","display_name":"Cloud computing","score":0.6324999928474426},{"id":"https://openalex.org/keywords/control-reconfiguration","display_name":"Control reconfiguration","score":0.5782999992370605},{"id":"https://openalex.org/keywords/controller","display_name":"Controller (irrigation)","score":0.5407000184059143},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5217000246047974},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.4510999917984009},{"id":"https://openalex.org/keywords/isolation","display_name":"Isolation (microbiology)","score":0.42730000615119934}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7670000195503235},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.7134000062942505},{"id":"https://openalex.org/C79974875","wikidata":"https://www.wikidata.org/wiki/Q483639","display_name":"Cloud computing","level":2,"score":0.6324999928474426},{"id":"https://openalex.org/C119701452","wikidata":"https://www.wikidata.org/wiki/Q5165881","display_name":"Control reconfiguration","level":2,"score":0.5782999992370605},{"id":"https://openalex.org/C203479927","wikidata":"https://www.wikidata.org/wiki/Q5165939","display_name":"Controller (irrigation)","level":2,"score":0.5407000184059143},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5217000246047974},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.46630001068115234},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.4510999917984009},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.43549999594688416},{"id":"https://openalex.org/C2775941552","wikidata":"https://www.wikidata.org/wiki/Q25212305","display_name":"Isolation (microbiology)","level":2,"score":0.42730000615119934},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.3882000148296356},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.38429999351501465},{"id":"https://openalex.org/C29202148","wikidata":"https://www.wikidata.org/wiki/Q287260","display_name":"Resource allocation","level":2,"score":0.37400001287460327},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.3544999957084656},{"id":"https://openalex.org/C32833848","wikidata":"https://www.wikidata.org/wiki/Q4115054","display_name":"Extensibility","level":2,"score":0.3059000074863434},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.28209999203681946},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.28139999508857727},{"id":"https://openalex.org/C100800780","wikidata":"https://www.wikidata.org/wiki/Q1175867","display_name":"Memory controller","level":3,"score":0.2632000148296356},{"id":"https://openalex.org/C206345919","wikidata":"https://www.wikidata.org/wiki/Q20380951","display_name":"Resource (disambiguation)","level":2,"score":0.26260000467300415}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3774899.3775014","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3774899.3775014","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 11th International Workshop on Serverless Computing","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":4,"referenced_works":["https://openalex.org/W2990215301","https://openalex.org/W4304192541","https://openalex.org/W4393904734","https://openalex.org/W4408886114"],"related_works":[],"abstract_inverted_index":{"Multi-tenant":[0],"LLM":[1,112],"inference":[2],"platforms":[3],"face":[4],"unpredictable":[5],"tail":[6,49],"latency":[7,50],"from":[8],"noisy-neighbor":[9],"interference,":[10,61],"violating":[11],"service-level":[12],"objectives":[13],"(SLOs)":[14],"critical":[15],"for":[16,53,110],"serverless":[17,111],"and":[18,41,62,95],"elastic":[19,107],"cloud":[20,71],"workloads.":[21],"We":[22],"present":[23],"a":[24,85],"VM-deployable":[25],"controller":[26,46],"that":[27],"enables":[28],"fine-grained,":[29],"dynamic":[30],"resource":[31],"allocation":[32,109],"by":[33,93,99],"combining":[34],"Multi-Instance":[35],"GPU":[36,108],"(MIG)":[37],"reconfiguration,":[38],"PCIe-aware":[39],"placement,":[40],"lightweight":[42],"isolation":[43,65],"guardrails.":[44],"The":[45],"samples":[47],"per-tenant":[48],"(including":[51],"time-to-first-token":[52],"autoregressive":[54],"serving),":[55],"correlates":[56],"system":[57],"signals":[58],"to":[59],"detect":[60],"adaptively":[63],"adjusts":[64],"using":[66],"host-only":[67],"controls":[68],"deployable":[69],"in":[70],"environments":[72],"without":[73],"fabric":[74],"privileges.":[75],"Evaluated":[76],"on":[77],"vLLM":[78],"serving":[79],"OLMo":[80],"2":[81],"7B":[82],"Instruct":[83],"across":[84],"16-GPU":[86],"cluster,":[87],"our":[88],"approach":[89],"reduces":[90],"SLO":[91],"miss-rate":[92],"\u224832%":[94],"improves":[96],"TTFT":[97],"p99":[98],"\u224810\u201315%":[100],"with":[101],"\u22645%":[102],"throughput":[103],"cost,":[104],"demonstrating":[105],"practical":[106],"inference.":[113]},"counts_by_year":[],"updated_date":"2025-12-11T00:21:10.989143","created_date":"2025-12-10T00:00:00"}