{"id":"https://openalex.org/W4394945061","doi":"https://doi.org/10.1145/3642970.3655830","title":"Characterizing Training Performance and Energy for Foundation Models and Image Classifiers on Multi-Instance GPUs","display_name":"Characterizing Training Performance and Energy for Foundation Models and Image Classifiers on Multi-Instance GPUs","publication_year":2024,"publication_date":"2024-04-19","ids":{"openalex":"https://openalex.org/W4394945061","doi":"https://doi.org/10.1145/3642970.3655830"},"language":"en","primary_location":{"id":"doi:10.1145/3642970.3655830","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3642970.3655830","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 4th Workshop on Machine Learning and Systems","raw_type":"proceedings-article"},"type":"conference-paper","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/A5021442818","display_name":"Connor Espenshade","orcid":"https://orcid.org/0000-0002-8318-2596"},"institutions":[{"id":"https://openalex.org/I78577930","display_name":"Columbia University","ror":"https://ror.org/00hj8s172","country_code":"US","type":"education","lineage":["https://openalex.org/I78577930"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Connor Espenshade","raw_affiliation_strings":["Columbia University, New York, New York, USA"],"raw_orcid":"https://orcid.org/0000-0002-8318-2596","affiliations":[{"raw_affiliation_string":"Columbia University, New York, New York, USA","institution_ids":["https://openalex.org/I78577930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037469191","display_name":"Ruoyu Peng","orcid":"https://orcid.org/0009-0003-2108-2156"},"institutions":[{"id":"https://openalex.org/I78577930","display_name":"Columbia University","ror":"https://ror.org/00hj8s172","country_code":"US","type":"education","lineage":["https://openalex.org/I78577930"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rachel Peng","raw_affiliation_strings":["Columbia University, New York, New York, USA"],"raw_orcid":"https://orcid.org/0009-0003-2108-2156","affiliations":[{"raw_affiliation_string":"Columbia University, New York, New York, USA","institution_ids":["https://openalex.org/I78577930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087166251","display_name":"Eumin Hong","orcid":"https://orcid.org/0000-0002-7547-7132"},"institutions":[{"id":"https://openalex.org/I78577930","display_name":"Columbia University","ror":"https://ror.org/00hj8s172","country_code":"US","type":"education","lineage":["https://openalex.org/I78577930"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Eumin Hong","raw_affiliation_strings":["Columbia University, New York, New York, USA"],"raw_orcid":"https://orcid.org/0000-0002-7547-7132","affiliations":[{"raw_affiliation_string":"Columbia University, New York, New York, USA","institution_ids":["https://openalex.org/I78577930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5095776651","display_name":"Max Calman","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Max Calman","raw_affiliation_strings":["IBM Research Yorktown Heights, New York, USA"],"raw_orcid":"https://orcid.org/0009-0003-3153-3600","affiliations":[{"raw_affiliation_string":"IBM Research Yorktown Heights, New York, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004237513","display_name":"Yue Zhu","orcid":"https://orcid.org/0009-0005-2207-1904"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yue Zhu","raw_affiliation_strings":["IBM Research Yorktown Heights, New York, USA"],"raw_orcid":"https://orcid.org/0009-0005-2207-1904","affiliations":[{"raw_affiliation_string":"IBM Research Yorktown Heights, New York, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5049875510","display_name":"Pritish R. Parida","orcid":"https://orcid.org/0000-0003-4722-7854"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pritish Parida","raw_affiliation_strings":["IBM Research Yorktown Heights, New York, USA"],"raw_orcid":"https://orcid.org/0000-0003-4722-7854","affiliations":[{"raw_affiliation_string":"IBM Research Yorktown Heights, New York, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023638218","display_name":"Eun Kyung Lee","orcid":"https://orcid.org/0000-0002-2335-4798"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Eun Kyung Lee","raw_affiliation_strings":["IBM Research Yorktown Heights, New York, USA"],"raw_orcid":"https://orcid.org/0000-0002-2335-4798","affiliations":[{"raw_affiliation_string":"IBM Research Yorktown Heights, New York, USA","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5086084861","display_name":"Martha A. Kim","orcid":"https://orcid.org/0000-0001-6243-5753"},"institutions":[{"id":"https://openalex.org/I78577930","display_name":"Columbia University","ror":"https://ror.org/00hj8s172","country_code":"US","type":"education","lineage":["https://openalex.org/I78577930"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Martha A. Kim","raw_affiliation_strings":["Columbia University, New York, New York, USA"],"raw_orcid":"https://orcid.org/0000-0001-6243-5753","affiliations":[{"raw_affiliation_string":"Columbia University, New York, New York, USA","institution_ids":["https://openalex.org/I78577930"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":10,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"47","last_page":"55"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9998999834060669,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9998999834060669,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9975000023841858,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9965999722480774,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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":0.8507860898971558},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.6775350570678711},{"id":"https://openalex.org/keywords/graphics","display_name":"Graphics","score":0.6672829389572144},{"id":"https://openalex.org/keywords/suite","display_name":"Suite","score":0.6606914401054382},{"id":"https://openalex.org/keywords/efficient-energy-use","display_name":"Efficient energy use","score":0.5308701992034912},{"id":"https://openalex.org/keywords/gpu-cluster","display_name":"GPU cluster","score":0.5291707515716553},{"id":"https://openalex.org/keywords/general-purpose-computing-on-graphics-processing-units","display_name":"General-purpose computing on graphics processing units","score":0.5244767069816589},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.5019855499267578},{"id":"https://openalex.org/keywords/supercomputer","display_name":"Supercomputer","score":0.5011885166168213},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.4719373881816864},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.4264848828315735},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.3731311559677124},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.34634092450141907},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.32774585485458374},{"id":"https://openalex.org/keywords/computer-graphics","display_name":"Computer graphics (images)","score":0.23987916111946106},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.14953488111495972}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8507860898971558},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.6775350570678711},{"id":"https://openalex.org/C21442007","wikidata":"https://www.wikidata.org/wiki/Q1027879","display_name":"Graphics","level":2,"score":0.6672829389572144},{"id":"https://openalex.org/C79581498","wikidata":"https://www.wikidata.org/wiki/Q1367530","display_name":"Suite","level":2,"score":0.6606914401054382},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.5308701992034912},{"id":"https://openalex.org/C2781335571","wikidata":"https://www.wikidata.org/wiki/Q2633544","display_name":"GPU cluster","level":3,"score":0.5291707515716553},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.5244767069816589},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.5019855499267578},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.5011885166168213},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.4719373881816864},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4264848828315735},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3731311559677124},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.34634092450141907},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.32774585485458374},{"id":"https://openalex.org/C121684516","wikidata":"https://www.wikidata.org/wiki/Q7600677","display_name":"Computer graphics (images)","level":1,"score":0.23987916111946106},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.14953488111495972},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.0},{"id":"https://openalex.org/C166957645","wikidata":"https://www.wikidata.org/wiki/Q23498","display_name":"Archaeology","level":1,"score":0.0},{"id":"https://openalex.org/C95457728","wikidata":"https://www.wikidata.org/wiki/Q309","display_name":"History","level":0,"score":0.0},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.0},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0},{"id":"https://openalex.org/C555944384","wikidata":"https://www.wikidata.org/wiki/Q249","display_name":"Wireless","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3642970.3655830","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3642970.3655830","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 4th Workshop on Machine Learning and Systems","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G5985404882","display_name":null,"funder_award_id":"ECCS-2227609","funder_id":"https://openalex.org/F4320323817","funder_display_name":"Universitas Brawijaya"}],"funders":[{"id":"https://openalex.org/F4320323817","display_name":"Universitas Brawijaya","ror":"https://ror.org/01wk3d929"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":9,"referenced_works":["https://openalex.org/W2108598243","https://openalex.org/W2183341477","https://openalex.org/W3021970402","https://openalex.org/W3041439392","https://openalex.org/W4289828024","https://openalex.org/W4304192541","https://openalex.org/W4362469997","https://openalex.org/W4394969374","https://openalex.org/W6780186301"],"related_works":["https://openalex.org/W1963859303","https://openalex.org/W2364044215","https://openalex.org/W2389600408","https://openalex.org/W240129890","https://openalex.org/W3048701459","https://openalex.org/W2149078538","https://openalex.org/W2080146221","https://openalex.org/W2370314112","https://openalex.org/W1912958759","https://openalex.org/W2792081825"],"abstract_inverted_index":{"GPUs":[0,21,86,139,158],"are":[1],"becoming":[2],"a":[3,51,88,110,145],"scarce":[4],"resource":[5],"in":[6,117],"high":[7,115],"demand,":[8],"as":[9],"many":[10],"teams":[11],"build":[12],"and":[13,33,93,104],"train":[14],"increasingly":[15],"advanced":[16],"artificial":[17],"intelligence":[18],"workloads.":[19],"As":[20],"become":[22],"more":[23,27],"performant,":[24],"they":[25],"consume":[26],"energy,":[28],"with":[29,87,121],"NVIDIA's":[30],"latest":[31,157],"A100":[32],"H100":[34],"graphics":[35],"cards":[36],"consuming":[37],"upwards":[38],"of":[39,41,90,144,155],"700W":[40],"power.":[42],"This":[43],"paper":[44,66],"characterizes":[45],"how":[46],"best":[47],"to":[48,55,76,99,152],"scale":[49],"down":[50],"large":[52,80,94],"modern":[53],"GPU":[54,147],"suite":[56],"workloads":[57,69,135],"that":[58,84,132,159],"cannot":[59],"fully":[60],"exploit":[61],"an":[62],"entire":[63],"GPU.":[64,112],"The":[65],"measures":[67],"six":[68],"from":[70,109],"14":[71],"million":[72],"parameter":[73,79],"image":[74],"classifiers":[75],"1.5":[77],"billion":[78],"language":[81],"models,":[82,120],"finding":[83],"partitioned":[85,138],"mix":[89],"small,":[91],"medium,":[92],"partitions":[95],"can":[96,140],"deliver":[97],"up":[98],"33%":[100],"less":[101,127],"energy":[102],"demand":[103],"9%":[105],"higher":[106],"training":[107,124],"throughput":[108],"single":[111,146],"We":[113],"found":[114],"potential":[116],"fine-tuning":[118],"existing":[119],"55%":[122],"faster":[123],"at":[125],"42%":[126],"energy.":[128],"Our":[129],"results":[130],"suggest":[131],"multiplexing":[133],"small":[134],"onto":[136],"spatially":[137],"improve":[141],"the":[142,156],"efficiency":[143],"while":[148],"giving":[149],"clients":[150],"access":[151],"smaller":[153],"slices":[154],"better":[160],"suits":[161],"their":[162],"job's":[163],"demands.":[164]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":7},{"year":2024,"cited_by_count":2}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}