{"id":"https://openalex.org/W4411092069","doi":"https://doi.org/10.1145/3713082.3730390","title":"Good things come in small packages: Should we build AI clusters with Lite-GPUs?","display_name":"Good things come in small packages: Should we build AI clusters with Lite-GPUs?","publication_year":2025,"publication_date":"2025-05-14","ids":{"openalex":"https://openalex.org/W4411092069","doi":"https://doi.org/10.1145/3713082.3730390"},"language":"en","primary_location":{"id":"doi:10.1145/3713082.3730390","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3713082.3730390","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3713082.3730390","source":null,"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Workshop on Hot Topics in Operating Systems","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3713082.3730390","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5030217737","display_name":"Burcu Canakci","orcid":"https://orcid.org/0000-0002-8406-7768"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Burcu Canakci","raw_affiliation_strings":["Microsoft Research, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0002-8406-7768","affiliations":[{"raw_affiliation_string":"Microsoft Research, United Kingdom","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100738705","display_name":"Junyi Liu","orcid":"https://orcid.org/0000-0002-4277-1802"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Junyi Liu","raw_affiliation_strings":["Microsoft Research, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0002-4277-1802","affiliations":[{"raw_affiliation_string":"Microsoft Research, United Kingdom","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113941029","display_name":"Xingbo Wu","orcid":"https://orcid.org/0000-0003-1649-2612"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Xingbo Wu","raw_affiliation_strings":["Microsoft Research, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0003-1649-2612","affiliations":[{"raw_affiliation_string":"Microsoft Research, United Kingdom","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064555295","display_name":"Nathana\u00ebl Cheriere","orcid":"https://orcid.org/0009-0008-5661-0007"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Nathana\u00ebl Cheriere","raw_affiliation_strings":["Microsoft Research, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0008-5661-0007","affiliations":[{"raw_affiliation_string":"Microsoft Research, United Kingdom","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087216360","display_name":"Paolo Costa","orcid":"https://orcid.org/0000-0003-1939-5690"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Paolo Costa","raw_affiliation_strings":["Microsoft Research, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0003-1939-5690","affiliations":[{"raw_affiliation_string":"Microsoft Research, United Kingdom","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5035654161","display_name":"Sergey Legtchenko","orcid":"https://orcid.org/0009-0001-5596-8962"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Sergey Legtchenko","raw_affiliation_strings":["Microsoft Research, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0001-5596-8962","affiliations":[{"raw_affiliation_string":"Microsoft Research, United Kingdom","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024912604","display_name":"Dushyanth Narayanan","orcid":"https://orcid.org/0009-0000-1194-2958"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Dushyanth Narayanan","raw_affiliation_strings":["Microsoft Research, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0000-1194-2958","affiliations":[{"raw_affiliation_string":"Microsoft Research, United Kingdom","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5033317994","display_name":"Antony Rowstron","orcid":"https://orcid.org/0009-0009-5936-6895"},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Antony Rowstron","raw_affiliation_strings":["Microsoft Research, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0009-5936-6895","affiliations":[{"raw_affiliation_string":"Microsoft Research, United Kingdom","institution_ids":["https://openalex.org/I4210164937"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I4210164937"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":6,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"127","last_page":"135"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9998999834060669,"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":0.9998999834060669,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.9997000098228455,"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/T10101","display_name":"Cloud Computing and Resource Management","score":0.9973000288009644,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6499195694923401},{"id":"https://openalex.org/keywords/internet-of-things","display_name":"Internet of Things","score":0.43409743905067444},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.40422767400741577},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.39038437604904175},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.3584096431732178},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.2535001039505005}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6499195694923401},{"id":"https://openalex.org/C81860439","wikidata":"https://www.wikidata.org/wiki/Q251212","display_name":"Internet of Things","level":2,"score":0.43409743905067444},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.40422767400741577},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.39038437604904175},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3584096431732178},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.2535001039505005}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3713082.3730390","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3713082.3730390","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3713082.3730390","source":null,"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Workshop on Hot Topics in Operating Systems","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3713082.3730390","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3713082.3730390","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3713082.3730390","source":null,"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Workshop on Hot Topics in Operating Systems","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4411092069.pdf","grobid_xml":"https://content.openalex.org/works/W4411092069.grobid-xml"},"referenced_works_count":20,"referenced_works":["https://openalex.org/W2002555321","https://openalex.org/W2061410814","https://openalex.org/W2108658640","https://openalex.org/W2734941459","https://openalex.org/W3046344355","https://openalex.org/W3170887803","https://openalex.org/W4281758439","https://openalex.org/W4281972408","https://openalex.org/W4310818142","https://openalex.org/W4327898871","https://openalex.org/W4380874786","https://openalex.org/W4381612905","https://openalex.org/W4392567268","https://openalex.org/W4394871754","https://openalex.org/W4394998727","https://openalex.org/W4401211672","https://openalex.org/W4401211704","https://openalex.org/W4404239615","https://openalex.org/W4404400711","https://openalex.org/W4404401018"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W4245926026","https://openalex.org/W4311097251","https://openalex.org/W2586548817","https://openalex.org/W2390279801","https://openalex.org/W2625093826","https://openalex.org/W4391913857","https://openalex.org/W2358668433"],"abstract_inverted_index":{"To":[0],"match":[1],"the":[2,36,60,82,113],"blooming":[3],"demand":[4],"of":[5,38,64,71,81,84,116],"generative":[6],"AI":[7,43,65,97],"workloads,":[8],"GPU":[9],"designers":[10],"have":[11],"so":[12],"far":[13],"been":[14],"trying":[15],"to":[16,58],"pack":[17],"more":[18,20],"and":[19,22,27,41,53,62,78,105,124,127,131,137],"compute":[21],"memory":[23],"into":[24],"single":[25],"complex":[26],"expensive":[28],"packages.":[29],"However,":[30],"there":[31],"is":[32],"growing":[33],"uncertainty":[34],"about":[35],"scalability":[37],"individual":[39],"GPUs":[40,47,73],"thus":[42],"clusters,":[44],"as":[45],"state-of-the-art":[46],"are":[48],"already":[49],"displaying":[50],"packaging,":[51],"yield,":[52,123],"cooling":[54],"limitations.":[55],"We":[56,87],"propose":[57],"rethink":[59],"design":[61],"scaling":[63],"clusters":[66,70],"through":[67,102],"efficiently-connected":[68],"large":[69],"Lite-GPUs,":[72],"with":[74],"single,":[75],"small":[76],"dies":[77],"a":[79],"fraction":[80],"capabilities":[83],"larger":[85],"GPUs.":[86],"think":[88],"recent":[89],"advances":[90],"in":[91],"co-packaged":[92],"optics":[93],"can":[94],"enable":[95],"distributing":[96],"workloads":[98],"onto":[99],"many":[100],"Lite-GPUs":[101,117],"high":[103],"bandwidth":[104],"efficient":[106],"communication.":[107],"In":[108],"this":[109],"paper,":[110],"we":[111],"present":[112],"key":[114],"benefits":[115],"on":[118],"manufacturing":[119],"cost,":[120],"blast":[121],"radius,":[122],"power":[125],"efficiency;":[126],"discuss":[128],"systems":[129],"opportunities":[130],"challenges":[132],"around":[133],"resource,":[134],"workload,":[135],"memory,":[136],"network":[138],"management.":[139]},"counts_by_year":[{"year":2026,"cited_by_count":4},{"year":2025,"cited_by_count":2}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}