{"id":"https://openalex.org/W4399794472","doi":"https://doi.org/10.1145/3659995.3660038","title":"Breaking the Memory Wall: A Study of I/O Patterns and GPU Memory Utilization for Hybrid CPU-GPU Offloaded Optimizers","display_name":"Breaking the Memory Wall: A Study of I/O Patterns and GPU Memory Utilization for Hybrid CPU-GPU Offloaded Optimizers","publication_year":2024,"publication_date":"2024-06-03","ids":{"openalex":"https://openalex.org/W4399794472","doi":"https://doi.org/10.1145/3659995.3660038"},"language":"en","primary_location":{"id":"doi:10.1145/3659995.3660038","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3659995.3660038","pdf_url":null,"source":null,"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 14th Workshop on AI and Scientific Computing at Scale using Flexible Computing Infrastructures","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3659995.3660038","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5049218913","display_name":"Avinash Maurya","orcid":"https://orcid.org/0000-0002-8200-0148"},"institutions":[{"id":"https://openalex.org/I1282105669","display_name":"Argonne National Laboratory","ror":"https://ror.org/05gvnxz63","country_code":"US","type":"facility","lineage":["https://openalex.org/I1282105669","https://openalex.org/I1330989302","https://openalex.org/I39565521","https://openalex.org/I40347166"]},{"id":"https://openalex.org/I155173764","display_name":"Rochester Institute of Technology","ror":"https://ror.org/00v4yb702","country_code":"US","type":"education","lineage":["https://openalex.org/I155173764"]},{"id":"https://openalex.org/I180949307","display_name":"Illinois Institute of Technology","ror":"https://ror.org/037t3ry66","country_code":"US","type":"education","lineage":["https://openalex.org/I180949307"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Avinash Maurya","raw_affiliation_strings":["Department of Computer Science, Rochester Institute of Technology, Rochester, New York, United States of America","Argonne National Laboratory Lemont, IL, USA","Illinois Institute of Technology Chicago, IL, USA"],"raw_orcid":"https://orcid.org/0000-0002-8200-0148","affiliations":[{"raw_affiliation_string":"Department of Computer Science, Rochester Institute of Technology, Rochester, New York, United States of America","institution_ids":["https://openalex.org/I155173764"]},{"raw_affiliation_string":"Argonne National Laboratory Lemont, IL, USA","institution_ids":["https://openalex.org/I1282105669"]},{"raw_affiliation_string":"Illinois Institute of Technology Chicago, IL, USA","institution_ids":["https://openalex.org/I180949307"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5106355306","display_name":"Jie Ye","orcid":null},"institutions":[{"id":"https://openalex.org/I1282105669","display_name":"Argonne National Laboratory","ror":"https://ror.org/05gvnxz63","country_code":"US","type":"facility","lineage":["https://openalex.org/I1282105669","https://openalex.org/I1330989302","https://openalex.org/I39565521","https://openalex.org/I40347166"]},{"id":"https://openalex.org/I155173764","display_name":"Rochester Institute of Technology","ror":"https://ror.org/00v4yb702","country_code":"US","type":"education","lineage":["https://openalex.org/I155173764"]},{"id":"https://openalex.org/I180949307","display_name":"Illinois Institute of Technology","ror":"https://ror.org/037t3ry66","country_code":"US","type":"education","lineage":["https://openalex.org/I180949307"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jie Ye","raw_affiliation_strings":["Illinois Institute of Technology, Chicago, Illinois, USA","Argonne National Laboratory Lemont, IL, USA","Rochester Institute of Technology Rochester, NY, USA"],"raw_orcid":"https://orcid.org/0009-0003-3985-7896","affiliations":[{"raw_affiliation_string":"Illinois Institute of Technology, Chicago, Illinois, USA","institution_ids":["https://openalex.org/I180949307"]},{"raw_affiliation_string":"Argonne National Laboratory Lemont, IL, USA","institution_ids":["https://openalex.org/I1282105669"]},{"raw_affiliation_string":"Rochester Institute of Technology Rochester, NY, USA","institution_ids":["https://openalex.org/I155173764"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054011637","display_name":"M. Mustafa Rafique","orcid":"https://orcid.org/0000-0002-5034-2880"},"institutions":[{"id":"https://openalex.org/I1282105669","display_name":"Argonne National Laboratory","ror":"https://ror.org/05gvnxz63","country_code":"US","type":"facility","lineage":["https://openalex.org/I1282105669","https://openalex.org/I1330989302","https://openalex.org/I39565521","https://openalex.org/I40347166"]},{"id":"https://openalex.org/I155173764","display_name":"Rochester Institute of Technology","ror":"https://ror.org/00v4yb702","country_code":"US","type":"education","lineage":["https://openalex.org/I155173764"]},{"id":"https://openalex.org/I180949307","display_name":"Illinois Institute of Technology","ror":"https://ror.org/037t3ry66","country_code":"US","type":"education","lineage":["https://openalex.org/I180949307"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"M. Mustafa Rafique","raw_affiliation_strings":["Department of Computer Science, Rochester Institute of Technology, Rochester, New York, United States of America","Argonne National Laboratory Lemont, IL, USA","Illinois Institute of Technology Chicago, IL, USA"],"raw_orcid":"https://orcid.org/0000-0002-5034-2880","affiliations":[{"raw_affiliation_string":"Department of Computer Science, Rochester Institute of Technology, Rochester, New York, United States of America","institution_ids":["https://openalex.org/I155173764"]},{"raw_affiliation_string":"Argonne National Laboratory Lemont, IL, USA","institution_ids":["https://openalex.org/I1282105669"]},{"raw_affiliation_string":"Illinois Institute of Technology Chicago, IL, USA","institution_ids":["https://openalex.org/I180949307"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5046613458","display_name":"Franck Cappello","orcid":"https://orcid.org/0000-0002-7890-3934"},"institutions":[{"id":"https://openalex.org/I1282105669","display_name":"Argonne National Laboratory","ror":"https://ror.org/05gvnxz63","country_code":"US","type":"facility","lineage":["https://openalex.org/I1282105669","https://openalex.org/I1330989302","https://openalex.org/I39565521","https://openalex.org/I40347166"]},{"id":"https://openalex.org/I155173764","display_name":"Rochester Institute of Technology","ror":"https://ror.org/00v4yb702","country_code":"US","type":"education","lineage":["https://openalex.org/I155173764"]},{"id":"https://openalex.org/I180949307","display_name":"Illinois Institute of Technology","ror":"https://ror.org/037t3ry66","country_code":"US","type":"education","lineage":["https://openalex.org/I180949307"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Franck Cappello","raw_affiliation_strings":["Argonne National Laboratory, Lemont, Illinois, USA","Illinois Institute of Technology Chicago, IL, USA","Rochester Institute of Technology Rochester, NY, USA"],"raw_orcid":"https://orcid.org/0000-0002-7890-3934","affiliations":[{"raw_affiliation_string":"Argonne National Laboratory, Lemont, Illinois, USA","institution_ids":["https://openalex.org/I1282105669"]},{"raw_affiliation_string":"Illinois Institute of Technology Chicago, IL, USA","institution_ids":["https://openalex.org/I180949307"]},{"raw_affiliation_string":"Rochester Institute of Technology Rochester, NY, USA","institution_ids":["https://openalex.org/I155173764"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5085745891","display_name":"Bogdan Nicolae","orcid":"https://orcid.org/0000-0002-0661-7509"},"institutions":[{"id":"https://openalex.org/I1282105669","display_name":"Argonne National Laboratory","ror":"https://ror.org/05gvnxz63","country_code":"US","type":"facility","lineage":["https://openalex.org/I1282105669","https://openalex.org/I1330989302","https://openalex.org/I39565521","https://openalex.org/I40347166"]},{"id":"https://openalex.org/I155173764","display_name":"Rochester Institute of Technology","ror":"https://ror.org/00v4yb702","country_code":"US","type":"education","lineage":["https://openalex.org/I155173764"]},{"id":"https://openalex.org/I180949307","display_name":"Illinois Institute of Technology","ror":"https://ror.org/037t3ry66","country_code":"US","type":"education","lineage":["https://openalex.org/I180949307"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bogdan Nicolae","raw_affiliation_strings":["Argonne National Laboratory, Lemont, Illinois, United States of America","Illinois Institute of Technology Chicago, IL, USA","Rochester Institute of Technology Rochester, NY, USA"],"raw_orcid":"https://orcid.org/0000-0002-0661-7509","affiliations":[{"raw_affiliation_string":"Argonne National Laboratory, Lemont, Illinois, United States of America","institution_ids":["https://openalex.org/I1282105669"]},{"raw_affiliation_string":"Illinois Institute of Technology Chicago, IL, USA","institution_ids":["https://openalex.org/I180949307"]},{"raw_affiliation_string":"Rochester Institute of Technology Rochester, NY, USA","institution_ids":["https://openalex.org/I155173764"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.832,"has_fulltext":true,"cited_by_count":7,"citation_normalized_percentile":{"value":0.89129079,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"9","last_page":"16"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9995999932289124,"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":0.9995999932289124,"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":0.9934999942779541,"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"}},{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.993399977684021,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8648215532302856},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6983397006988525},{"id":"https://openalex.org/keywords/memory-management","display_name":"Memory management","score":0.4616829752922058},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.437161922454834},{"id":"https://openalex.org/keywords/host","display_name":"Host (biology)","score":0.4351620674133301},{"id":"https://openalex.org/keywords/general-purpose-computing-on-graphics-processing-units","display_name":"General-purpose computing on graphics processing units","score":0.4139166474342346},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.26293039321899414},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.20634281635284424},{"id":"https://openalex.org/keywords/semiconductor-memory","display_name":"Semiconductor memory","score":0.1329035460948944},{"id":"https://openalex.org/keywords/graphics","display_name":"Graphics","score":0.13148698210716248},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.11150279641151428}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8648215532302856},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6983397006988525},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.4616829752922058},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.437161922454834},{"id":"https://openalex.org/C126831891","wikidata":"https://www.wikidata.org/wiki/Q221673","display_name":"Host (biology)","level":2,"score":0.4351620674133301},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.4139166474342346},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.26293039321899414},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.20634281635284424},{"id":"https://openalex.org/C98986596","wikidata":"https://www.wikidata.org/wiki/Q1143031","display_name":"Semiconductor memory","level":2,"score":0.1329035460948944},{"id":"https://openalex.org/C21442007","wikidata":"https://www.wikidata.org/wiki/Q1027879","display_name":"Graphics","level":2,"score":0.13148698210716248},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.11150279641151428},{"id":"https://openalex.org/C18903297","wikidata":"https://www.wikidata.org/wiki/Q7150","display_name":"Ecology","level":1,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3659995.3660038","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3659995.3660038","pdf_url":null,"source":null,"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 14th Workshop on AI and Scientific Computing at Scale using Flexible Computing Infrastructures","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2406.10728","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2406.10728","pdf_url":"https://arxiv.org/pdf/2406.10728","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"doi:10.1145/3659995.3660038","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3659995.3660038","pdf_url":null,"source":null,"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 14th Workshop on AI and Scientific Computing at Scale using Flexible Computing Infrastructures","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G102163711","display_name":null,"funder_award_id":"DE\u2010 AC02\u201006CH11357","funder_id":"https://openalex.org/F4320306084","funder_display_name":"U.S. Department of Energy"},{"id":"https://openalex.org/G107546177","display_name":null,"funder_award_id":"DE- AC02-06-CH11357","funder_id":"https://openalex.org/F4320338284","funder_display_name":"Argonne National Laboratory"},{"id":"https://openalex.org/G2131453672","display_name":"Collaborative Research: CNS Core: Medium:HardLambda: A new FaaS Abstraction for Cross-Stack Resource Management in Disaggregated Datacenters","funder_award_id":"2106635","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G3075337988","display_name":null,"funder_award_id":"06CH11357","funder_id":"https://openalex.org/F4320338284","funder_display_name":"Argonne National Laboratory"},{"id":"https://openalex.org/G3085353284","display_name":null,"funder_award_id":"2106634\\/2106635","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G4552540348","display_name":null,"funder_award_id":"0F-60169","funder_id":"https://openalex.org/F4320306084","funder_display_name":"U.S. Department of Energy"},{"id":"https://openalex.org/G5367729914","display_name":null,"funder_award_id":"0F-60169","funder_id":"https://openalex.org/F4320338284","funder_display_name":"Argonne National Laboratory"},{"id":"https://openalex.org/G6848031779","display_name":null,"funder_award_id":"06CH11357","funder_id":"https://openalex.org/F4320306084","funder_display_name":"U.S. Department of Energy"},{"id":"https://openalex.org/G6918803902","display_name":null,"funder_award_id":"06CH11357","funder_id":"https://openalex.org/F4320332359","funder_display_name":"Office of Science"},{"id":"https://openalex.org/G8515454090","display_name":"Collaborative Research: CNS Core: Medium:HardLambda: A new FaaS Abstraction for Cross-Stack Resource Management in Disaggregated Datacenters","funder_award_id":"2106634","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G878601127","display_name":null,"funder_award_id":"DE-AC02-06CH11357","funder_id":"https://openalex.org/F4320337506","funder_display_name":"Advanced Scientific Computing Research"},{"id":"https://openalex.org/G993149817","display_name":null,"funder_award_id":"DE-AC02-06CH11357 ","funder_id":"https://openalex.org/F4320332359","funder_display_name":"Office of Science"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320306084","display_name":"U.S. Department of Energy","ror":"https://ror.org/01bj3aw27"},{"id":"https://openalex.org/F4320332359","display_name":"Office of Science","ror":"https://ror.org/00mmn6b08"},{"id":"https://openalex.org/F4320337506","display_name":"Advanced Scientific Computing Research","ror":"https://ror.org/0012c7r22"},{"id":"https://openalex.org/F4320338284","display_name":"Argonne National Laboratory","ror":"https://ror.org/05gvnxz63"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":17,"referenced_works":["https://openalex.org/W2973727699","https://openalex.org/W3025935268","https://openalex.org/W3086105743","https://openalex.org/W3119866685","https://openalex.org/W3121562065","https://openalex.org/W3156643189","https://openalex.org/W3206606249","https://openalex.org/W4280611847","https://openalex.org/W4287391717","https://openalex.org/W4308760226","https://openalex.org/W4376226279","https://openalex.org/W4384918448","https://openalex.org/W4386707654","https://openalex.org/W4387321091","https://openalex.org/W4387559225","https://openalex.org/W6846002521","https://openalex.org/W6851775633"],"related_works":["https://openalex.org/W2505380084","https://openalex.org/W4400333498","https://openalex.org/W2086739451","https://openalex.org/W3183233360","https://openalex.org/W2073137634","https://openalex.org/W2808906329","https://openalex.org/W22188748","https://openalex.org/W3019930626","https://openalex.org/W2117112636","https://openalex.org/W1980160788"],"abstract_inverted_index":{"Transformers":[0],"and":[1,20,32,49,101,118,155,177,183,229,233,239,273],"LLMs":[2],"have":[3,12],"seen":[4],"rapid":[5],"adoption":[6],"in":[7,35,169,243,268],"all":[8],"domains.":[9],"Their":[10],"sizes":[11],"exploded":[13],"to":[14,43,56,91,114,136,190,223,247,263],"hundreds":[15],"of":[16,18,28,38,60,66,79,144,149,157,172,198,256,271],"billions":[17],"parameters":[19],"keep":[21],"increasing.":[22],"Under":[23],"these":[24],"circumstances,":[25],"the":[26,36,52,64,67,72,83,93,108,115,128,138,147,153,173,196,202,208,217,220,226,230,234,240,265,269],"training":[27,53,68,95,139,200],"transformers":[29],"is":[30,87],"slow":[31],"often":[33,88],"takes":[34],"order":[37],"weeks":[39],"or":[40],"months.":[41],"Thanks":[42,246],"3D":[44],"model":[45,99],"parallelism":[46],"(data,":[47],"pipeline,":[48],"tensor-level":[50],"parallelism),":[51],"can":[54],"scale":[55],"a":[57,76,141],"large":[58,77],"number":[59,78,143],"GPUs,":[61],"which":[62,132,167,259],"reduces":[63],"duration":[65],"but":[69],"dramatically":[70,126],"increases":[71],"cost.":[73],"Even":[74],"when":[75],"GPUs":[80,145],"are":[81,161],"available,":[82],"aggregated":[84],"GPU":[85,129,209,231],"memory":[86,117,130,176,210,228],"not":[89,162],"enough":[90],"hold":[92],"full":[94],"state":[96,110],"(optimizer":[97],"state,":[98],"parameters,":[100],"gradients).":[102],"To":[103],"compensate,":[104],"state-of-the-art":[105],"approaches":[106],"offload":[107],"optimizer":[109],"at":[111,146],"least":[112],"partially":[113],"host":[116,227],"perform":[119],"hybrid":[120],"CPU-GPU":[121],"computations.":[122,184],"Such":[123],"flexible":[124],"solutions":[125],"reduce":[127],"utilization,":[131],"makes":[133],"it":[134],"feasible":[135],"run":[137],"on":[140,219],"smaller":[142],"cost":[148],"performance":[150],"penalty.":[151],"Unfortunately,":[152],"challenges":[154],"bottlenecks":[156],"adopting":[158],"this":[159,186,192,248],"strategy":[160],"sufficiently":[163],"studied":[164],"by":[165,194],"state-of-the-art,":[166],"results":[168],"poor":[170,178],"management":[171],"combined":[174],"host-GPU":[175],"overlapping":[179],"between":[180,225,236],"data":[181],"movements":[182],"In":[185],"paper,":[187],"we":[188,206,250],"aim":[189],"fill":[191],"gap":[193],"characterizing":[195],"behavior":[197],"offloaded":[199],"using":[201],"DeepSpeed":[203],"runtime.":[204],"Specifically,":[205],"study":[207],"utilization":[211,238],"over":[212],"time":[213],"during":[214],"each":[215,244],"iteration,":[216],"activity":[218],"PCIe":[221],"related":[222],"transfers":[224],"memory,":[232],"relationship":[235],"resource":[237],"steps":[241],"involved":[242],"iteration.":[245],"study,":[249],"reveal":[251],"opportunities":[252],"for":[253],"future":[254],"improvements":[255],"offloading":[257],"solutions,":[258],"enable":[260],"greater":[261],"flexibility":[262],"optimize":[264],"cost-performance":[266],"trade-off":[267],"context":[270],"transformer":[272],"LLM":[274],"training.":[275]},"counts_by_year":[{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":1}],"updated_date":"2026-08-01T09:00:35.917206","created_date":"2024-06-19T00:00:00"}
