{"id":"https://openalex.org/W4280602165","doi":"https://doi.org/10.1109/hpca53966.2022.00018","title":"Enabling Efficient Large-Scale Deep Learning Training with Cache Coherent Disaggregated Memory Systems","display_name":"Enabling Efficient Large-Scale Deep Learning Training with Cache Coherent Disaggregated Memory Systems","publication_year":2022,"publication_date":"2022-04-01","ids":{"openalex":"https://openalex.org/W4280602165","doi":"https://doi.org/10.1109/hpca53966.2022.00018"},"language":"en","primary_location":{"id":"doi:10.1109/hpca53966.2022.00018","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpca53966.2022.00018","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA)","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/A5100398240","display_name":"Zixuan Wang","orcid":"https://orcid.org/0000-0001-8677-7993"},"institutions":[{"id":"https://openalex.org/I36258959","display_name":"University of California, San Diego","ror":"https://ror.org/0168r3w48","country_code":"US","type":"education","lineage":["https://openalex.org/I36258959"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Zixuan Wang","raw_affiliation_strings":["University of California,San Diego","University of California, San Diego"],"affiliations":[{"raw_affiliation_string":"University of California,San Diego","institution_ids":["https://openalex.org/I36258959"]},{"raw_affiliation_string":"University of California, San Diego","institution_ids":["https://openalex.org/I36258959"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5053502690","display_name":"Joonseop Sim","orcid":"https://orcid.org/0000-0003-3417-0289"},"institutions":[{"id":"https://openalex.org/I36258959","display_name":"University of California, San Diego","ror":"https://ror.org/0168r3w48","country_code":"US","type":"education","lineage":["https://openalex.org/I36258959"]},{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Joonseop Sim","raw_affiliation_strings":["University of California,San Diego","University of California, San Diego","System Architecture Division, SK Hynix"],"affiliations":[{"raw_affiliation_string":"University of California,San Diego","institution_ids":["https://openalex.org/I36258959"]},{"raw_affiliation_string":"University of California, San Diego","institution_ids":["https://openalex.org/I36258959"]},{"raw_affiliation_string":"System Architecture Division, SK Hynix","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039338759","display_name":"Euicheol Lim","orcid":"https://orcid.org/0000-0002-8910-533X"},"institutions":[{"id":"https://openalex.org/I36258959","display_name":"University of California, San Diego","ror":"https://ror.org/0168r3w48","country_code":"US","type":"education","lineage":["https://openalex.org/I36258959"]},{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Euicheol Lim","raw_affiliation_strings":["University of California,San Diego","System Architecture Division, SK Hynix","University of California, San Diego"],"affiliations":[{"raw_affiliation_string":"University of California,San Diego","institution_ids":["https://openalex.org/I36258959"]},{"raw_affiliation_string":"System Architecture Division, SK Hynix","institution_ids":["https://openalex.org/I10654025"]},{"raw_affiliation_string":"University of California, San Diego","institution_ids":["https://openalex.org/I36258959"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5077387335","display_name":"Jishen Zhao","orcid":"https://orcid.org/0000-0002-1969-743X"},"institutions":[{"id":"https://openalex.org/I36258959","display_name":"University of California, San Diego","ror":"https://ror.org/0168r3w48","country_code":"US","type":"education","lineage":["https://openalex.org/I36258959"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jishen Zhao","raw_affiliation_strings":["University of California,San Diego","University of California, San Diego"],"affiliations":[{"raw_affiliation_string":"University of California,San Diego","institution_ids":["https://openalex.org/I36258959"]},{"raw_affiliation_string":"University of California, San Diego","institution_ids":["https://openalex.org/I36258959"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5100398240"],"corresponding_institution_ids":["https://openalex.org/I36258959"],"apc_list":null,"apc_paid":null,"fwci":0.8233,"has_fulltext":false,"cited_by_count":9,"citation_normalized_percentile":{"value":0.70320666,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"126","last_page":"140"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9993000030517578,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9993000030517578,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.9986000061035156,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9984999895095825,"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":0.8765636086463928},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.690260648727417},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.5801404118537903},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.5683795213699341},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.5441861748695374},{"id":"https://openalex.org/keywords/synchronization","display_name":"Synchronization (alternating current)","score":0.5149134397506714},{"id":"https://openalex.org/keywords/bandwidth","display_name":"Bandwidth (computing)","score":0.49613437056541443},{"id":"https://openalex.org/keywords/cache-coherence","display_name":"Cache coherence","score":0.4670765697956085},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.43080461025238037},{"id":"https://openalex.org/keywords/memory-bandwidth","display_name":"Memory bandwidth","score":0.4266061782836914},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4111233353614807},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.27185142040252686},{"id":"https://openalex.org/keywords/cpu-cache","display_name":"CPU cache","score":0.25449180603027344},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.24700099229812622},{"id":"https://openalex.org/keywords/cache-algorithms","display_name":"Cache algorithms","score":0.14357823133468628},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.09132802486419678}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8765636086463928},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.690260648727417},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5801404118537903},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.5683795213699341},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5441861748695374},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.5149134397506714},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.49613437056541443},{"id":"https://openalex.org/C141917322","wikidata":"https://www.wikidata.org/wiki/Q1025017","display_name":"Cache coherence","level":5,"score":0.4670765697956085},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.43080461025238037},{"id":"https://openalex.org/C188045654","wikidata":"https://www.wikidata.org/wiki/Q17148339","display_name":"Memory bandwidth","level":2,"score":0.4266061782836914},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4111233353614807},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.27185142040252686},{"id":"https://openalex.org/C189783530","wikidata":"https://www.wikidata.org/wiki/Q352090","display_name":"CPU cache","level":3,"score":0.25449180603027344},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.24700099229812622},{"id":"https://openalex.org/C38556500","wikidata":"https://www.wikidata.org/wiki/Q13404475","display_name":"Cache algorithms","level":4,"score":0.14357823133468628},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.09132802486419678},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpca53966.2022.00018","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpca53966.2022.00018","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":54,"referenced_works":["https://openalex.org/W1442374986","https://openalex.org/W1970663544","https://openalex.org/W2007511350","https://openalex.org/W2060393849","https://openalex.org/W2108598243","https://openalex.org/W2117539524","https://openalex.org/W2132424367","https://openalex.org/W2156094106","https://openalex.org/W2160815625","https://openalex.org/W2161047342","https://openalex.org/W2168231600","https://openalex.org/W2302255633","https://openalex.org/W2339765813","https://openalex.org/W2550821151","https://openalex.org/W2617411258","https://openalex.org/W2775461895","https://openalex.org/W2786414509","https://openalex.org/W2794670651","https://openalex.org/W2799610839","https://openalex.org/W2896457183","https://openalex.org/W2903754802","https://openalex.org/W2911382970","https://openalex.org/W2919115771","https://openalex.org/W2926767350","https://openalex.org/W2949161920","https://openalex.org/W2953212265","https://openalex.org/W2953384591","https://openalex.org/W2963748441","https://openalex.org/W2975712713","https://openalex.org/W2979719709","https://openalex.org/W2981587687","https://openalex.org/W2991040477","https://openalex.org/W3010969086","https://openalex.org/W3011751313","https://openalex.org/W3037847693","https://openalex.org/W3043522163","https://openalex.org/W3101708369","https://openalex.org/W4287774547","https://openalex.org/W4288093768","https://openalex.org/W4289401659","https://openalex.org/W4301239768","https://openalex.org/W6628377381","https://openalex.org/W6684859321","https://openalex.org/W6713134421","https://openalex.org/W6738144653","https://openalex.org/W6746932923","https://openalex.org/W6748645090","https://openalex.org/W6748687944","https://openalex.org/W6753584990","https://openalex.org/W6755207826","https://openalex.org/W6756718674","https://openalex.org/W6769475105","https://openalex.org/W6774806506","https://openalex.org/W6778729859"],"related_works":["https://openalex.org/W2469267785","https://openalex.org/W3133724979","https://openalex.org/W1997955449","https://openalex.org/W2113103358","https://openalex.org/W2048574738","https://openalex.org/W1588772634","https://openalex.org/W2131630752","https://openalex.org/W1503860218","https://openalex.org/W1579918296","https://openalex.org/W1506505813"],"abstract_inverted_index":{"Modern":[0],"deep":[1],"learning":[2],"(DL)":[3],"training":[4,32,89,171],"is":[5,41,68],"memory-consuming,":[6],"constrained":[7],"by":[8],"the":[9,106,135,174],"memory":[10,61,108],"capacity":[11],"of":[12],"each":[13],"computation":[14],"component":[15],"and":[16,33,76,85,91,105,119,130,152],"cross-device":[17],"communication":[18,40,54,79,115],"bandwidth.":[19],"In":[20],"response":[21],"to":[22,82,88,117,132,155,167,173],"such":[23],"constraints,":[24],"current":[25],"approaches":[26],"include":[27],"increasing":[28],"parallelism":[29],"in":[30,47],"distributed":[31,48,64],"optimizing":[34],"inter-device":[35],"communication.":[36,178],"However,":[37],"model":[38,92],"parameter":[39,53,114,121,158],"becoming":[42],"a":[43,59,112,149],"key":[44],"performance":[45],"bottleneck":[46],"DL":[49,65,170],"training.":[50,66],"To":[51,98],"improve":[52],"performance,":[55],"we":[56,110,125,147],"propose":[57,111,126],"COARSE,":[58],"disaggregated":[60,107],"extension":[62],"for":[63,80],"COARSE":[67,164],"built":[69],"on":[70],"modern":[71],"cache-coherent":[72],"interconnect":[73],"(CCI)":[74],"protocols":[75],"MPI-like":[77],"collective":[78],"synchronization,":[81],"allow":[83],"low-latency":[84],"parallel":[86],"access":[87],"data":[90],"parameters":[93],"shared":[94],"among":[95],"worker":[96],"GPUs.":[97],"enable":[99],"high":[100],"bandwidth":[101,139],"transfers":[102],"between":[103],"GPUs":[104],"system,":[109],"decentralized":[113],"scheme":[116],"decouple":[118],"localize":[120],"synchronization":[122,154],"traffic.":[123],"Furthermore,":[124],"dynamic":[127],"tensor":[128],"routing":[129],"partitioning":[131],"fully":[133],"utilize":[134],"non-uniform":[136],"serial":[137],"bus":[138],"varied":[140],"across":[141],"different":[142],"cloud":[143],"computing":[144],"systems.":[145],"Finally,":[146],"design":[148],"deadlock":[150],"avoidance":[151],"dual":[153],"ensure":[156],"high-performance":[157],"synchronization.":[159],"Our":[160],"evaluation":[161],"shows":[162],"that":[163],"achieves":[165],"up":[166],"48.3%":[168],"faster":[169],"compared":[172],"state-of-the-art":[175],"MPI":[176],"AllReduce":[177]},"counts_by_year":[{"year":2025,"cited_by_count":1},{"year":2024,"cited_by_count":3},{"year":2023,"cited_by_count":2},{"year":2022,"cited_by_count":3}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}