{"id":"https://openalex.org/W3206003350","doi":"https://doi.org/10.1145/3466752.3480088","title":"GPS: A Global Publish-Subscribe Model for Multi-GPU Memory Management","display_name":"GPS: A Global Publish-Subscribe Model for Multi-GPU Memory Management","publication_year":2021,"publication_date":"2021-10-17","ids":{"openalex":"https://openalex.org/W3206003350","doi":"https://doi.org/10.1145/3466752.3480088","mag":"3206003350"},"language":"en","primary_location":{"id":"doi:10.1145/3466752.3480088","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3466752.3480088","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"MICRO-54: 54th Annual IEEE/ACM International Symposium on Microarchitecture","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/A5006855402","display_name":"Harini Muthukrishnan","orcid":"https://orcid.org/0009-0001-7938-9623"},"institutions":[{"id":"https://openalex.org/I27837315","display_name":"University of Michigan","ror":"https://ror.org/00jmfr291","country_code":"US","type":"education","lineage":["https://openalex.org/I27837315"]},{"id":"https://openalex.org/I4210111179","display_name":"Michigan United","ror":"https://ror.org/0291ys696","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I4210111179"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Harini Muthukrishnan","raw_affiliation_strings":["University of Michigan, United States of America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Michigan, United States of America","institution_ids":["https://openalex.org/I27837315","https://openalex.org/I4210111179"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078894707","display_name":"Daniel Lustig","orcid":"https://orcid.org/0000-0001-9763-7304"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Daniel Lustig","raw_affiliation_strings":["NVIDIA, United States of America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA, United States of America","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031781240","display_name":"David Nellans","orcid":"https://orcid.org/0000-0001-5203-8367"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"David Nellans","raw_affiliation_strings":["NVIDIA, United States of America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA, United States of America","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5018949021","display_name":"Thomas F. Wenisch","orcid":"https://orcid.org/0000-0001-9560-2124"},"institutions":[{"id":"https://openalex.org/I27837315","display_name":"University of Michigan","ror":"https://ror.org/00jmfr291","country_code":"US","type":"education","lineage":["https://openalex.org/I27837315"]},{"id":"https://openalex.org/I4210111179","display_name":"Michigan United","ror":"https://ror.org/0291ys696","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I4210111179"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Thomas Wenisch","raw_affiliation_strings":["University of Michigan, United States of America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Michigan, United States of America","institution_ids":["https://openalex.org/I27837315","https://openalex.org/I4210111179"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.0844,"has_fulltext":false,"cited_by_count":19,"citation_normalized_percentile":{"value":0.87639292,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":96,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"46","last_page":"58"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.9998999834060669,"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":0.9998999834060669,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9994000196456909,"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/T11478","display_name":"Caching and Content Delivery","score":0.9991999864578247,"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/computer-science","display_name":"Computer science","score":0.883462131023407},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.7255787253379822},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.5715221166610718},{"id":"https://openalex.org/keywords/shared-memory","display_name":"Shared memory","score":0.5089385509490967},{"id":"https://openalex.org/keywords/global-positioning-system","display_name":"Global Positioning System","score":0.505396842956543},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.4687158167362213},{"id":"https://openalex.org/keywords/general-purpose-computing-on-graphics-processing-units","display_name":"General-purpose computing on graphics processing units","score":0.4649343192577362},{"id":"https://openalex.org/keywords/memory-management","display_name":"Memory management","score":0.45936933159828186},{"id":"https://openalex.org/keywords/programming-paradigm","display_name":"Programming paradigm","score":0.43571770191192627},{"id":"https://openalex.org/keywords/memory-model","display_name":"Memory model","score":0.43133246898651123},{"id":"https://openalex.org/keywords/memory-bandwidth","display_name":"Memory bandwidth","score":0.4212993085384369},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.19763553142547607},{"id":"https://openalex.org/keywords/graphics","display_name":"Graphics","score":0.1753406822681427},{"id":"https://openalex.org/keywords/semiconductor-memory","display_name":"Semiconductor memory","score":0.12902125716209412}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.883462131023407},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.7255787253379822},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5715221166610718},{"id":"https://openalex.org/C133875982","wikidata":"https://www.wikidata.org/wiki/Q764810","display_name":"Shared memory","level":2,"score":0.5089385509490967},{"id":"https://openalex.org/C60229501","wikidata":"https://www.wikidata.org/wiki/Q18822","display_name":"Global Positioning System","level":2,"score":0.505396842956543},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.4687158167362213},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.4649343192577362},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.45936933159828186},{"id":"https://openalex.org/C34165917","wikidata":"https://www.wikidata.org/wiki/Q188267","display_name":"Programming paradigm","level":2,"score":0.43571770191192627},{"id":"https://openalex.org/C12186640","wikidata":"https://www.wikidata.org/wiki/Q6815743","display_name":"Memory model","level":3,"score":0.43133246898651123},{"id":"https://openalex.org/C188045654","wikidata":"https://www.wikidata.org/wiki/Q17148339","display_name":"Memory bandwidth","level":2,"score":0.4212993085384369},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.19763553142547607},{"id":"https://openalex.org/C21442007","wikidata":"https://www.wikidata.org/wiki/Q1027879","display_name":"Graphics","level":2,"score":0.1753406822681427},{"id":"https://openalex.org/C98986596","wikidata":"https://www.wikidata.org/wiki/Q1143031","display_name":"Semiconductor memory","level":2,"score":0.12902125716209412},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3466752.3480088","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3466752.3480088","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"MICRO-54: 54th Annual IEEE/ACM International Symposium on Microarchitecture","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":25,"referenced_works":["https://openalex.org/W1817506983","https://openalex.org/W1979717209","https://openalex.org/W2053776346","https://openalex.org/W2062430565","https://openalex.org/W2067580069","https://openalex.org/W2094191591","https://openalex.org/W2098931293","https://openalex.org/W2135393198","https://openalex.org/W2157802978","https://openalex.org/W2159184943","https://openalex.org/W2233147800","https://openalex.org/W2236227338","https://openalex.org/W2245726859","https://openalex.org/W2401249903","https://openalex.org/W2613380095","https://openalex.org/W2625200202","https://openalex.org/W2761710529","https://openalex.org/W2903659818","https://openalex.org/W2903901007","https://openalex.org/W2935389012","https://openalex.org/W2951135776","https://openalex.org/W2979340153","https://openalex.org/W3036082746","https://openalex.org/W3157055696","https://openalex.org/W3189320660"],"related_works":["https://openalex.org/W2349791858","https://openalex.org/W2007942862","https://openalex.org/W2311112479","https://openalex.org/W4317827391","https://openalex.org/W2033109261","https://openalex.org/W40577698","https://openalex.org/W1578381328","https://openalex.org/W1839004754","https://openalex.org/W2497938157","https://openalex.org/W4233212676"],"abstract_inverted_index":{"Suboptimal":[0],"management":[1,22,67,190],"of":[2,8,12,36,83,90,110,143,157,172,179,221],"memory":[3,21,66,86,136,189],"and":[4,118,164],"bandwidth":[5],"is":[6,129],"one":[7],"the":[9,34,80,88,99,122,132,155,177,184,222],"primary":[10],"causes":[11],"low":[13],"performance":[14,89,178],"on":[15,195],"systems":[16],"comprising":[17],"multiple":[18],"GPUs.":[19,127],"Existing":[20],"solutions":[23],"like":[24],"Unified":[25],"Memory":[26],"(UM)":[27],"offer":[28],"simplified":[29],"programming":[30],"but":[31,139],"come":[32],"at":[33],"cost":[35],"performance:":[37],"applications":[38],"can":[39],"even":[40],"exhibit":[41],"slowdown":[42],"with":[43,87,161],"increasing":[44],"GPU":[45,135],"count":[46],"due":[47],"to":[48,51,121,147,176],"their":[49],"inability":[50],"leverage":[52],"system":[53,160],"resources":[54],"effectively.":[55],"To":[56,93],"solve":[57],"this":[58],"challenge,":[59],"we":[60,207],"propose":[61],"GPS,":[62],"a":[63,158,180,198,202,209],"HW/SW":[64],"multi-GPU":[65,84,188],"technique":[68,191],"that":[69,166],"efficiently":[70],"orchestrates":[71],"inter-GPU":[72],"communication":[73],"using":[74,201],"proactive":[75],"data":[76,100],"transfers.":[77],"GPS":[78,96,128,153,167],"offers":[79],"programmability":[81],"advantage":[82,142],"shared":[85,111],"GPU-local":[91],"memory.":[92],"enable":[94],"this,":[95],"automatically":[97],"tracks":[98],"accesses":[101],"performed":[102],"by":[103,192],"each":[104,114],"GPU,":[105,182],"maintains":[106],"duplicate":[107],"physical":[108],"replicas":[109,123],"regions":[112],"in":[113,124,154],"GPU\u2019s":[115],"local":[116],"memory,":[117],"pushes":[119],"updates":[120],"all":[125],"consumer":[126],"compatible":[130],"within":[131],"existing":[133],"NVIDIA":[134],"consistency":[137],"model":[138],"takes":[140],"full":[141],"its":[144],"relaxed":[145],"nature":[146],"deliver":[148],"high":[149],"performance.":[150],"We":[151],"evaluate":[152],"context":[156],"4-GPU":[159],"varying":[162],"interconnects":[163],"show":[165],"achieves":[168],"an":[169],"average":[170,212],"speedup":[171,215],"3.0":[173],"\u00d7":[174,194,211],"relative":[175],"single":[181],"outperforming":[183],"next":[185],"best":[186],"available":[187,223],"2.3":[193],"average.":[196],"In":[197],"16-GPU":[199],"system,":[200],"future":[203],"PCIe":[204],"6.0":[205],"interconnect,":[206],"demonstrate":[208],"7.9":[210],"strong":[213],"scaling":[214],"over":[216],"single-GPU":[217],"performance,":[218],"capturing":[219],"80%":[220],"opportunity.":[224]},"counts_by_year":[{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":5},{"year":2023,"cited_by_count":5},{"year":2022,"cited_by_count":3}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}