{"id":"https://openalex.org/W4224952058","doi":"https://doi.org/10.1109/lca.2022.3189207","title":"FPGA-Based AI Smart NICs for Scalable Distributed AI Training Systems","display_name":"FPGA-Based AI Smart NICs for Scalable Distributed AI Training Systems","publication_year":2022,"publication_date":"2022-07-01","ids":{"openalex":"https://openalex.org/W4224952058","doi":"https://doi.org/10.1109/lca.2022.3189207"},"language":"en","primary_location":{"id":"doi:10.1109/lca.2022.3189207","is_oa":false,"landing_page_url":"https://doi.org/10.1109/lca.2022.3189207","pdf_url":null,"source":{"id":"https://openalex.org/S17643076","display_name":"IEEE Computer Architecture Letters","issn_l":"1556-6056","issn":["1556-6056","1556-6064","2473-2575"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Computer Architecture Letters","raw_type":"journal-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/A5084165072","display_name":"Rui Ma","orcid":"https://orcid.org/0000-0001-9611-5870"},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rui Ma","raw_affiliation_strings":["Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA"],"raw_orcid":"https://orcid.org/0000-0001-9611-5870","affiliations":[{"raw_affiliation_string":"Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057664789","display_name":"Evangelos Georganas","orcid":null},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Evangelos Georganas","raw_affiliation_strings":["Intel Corp., Santa Clara, CA, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Intel Corp., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024066643","display_name":"Alexander Heinecke","orcid":null},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Alexander Heinecke","raw_affiliation_strings":["Intel Corp., Santa Clara, CA, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Intel Corp., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060878942","display_name":"Sergey Gribok","orcid":"https://orcid.org/0000-0003-3339-7705"},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sergey Gribok","raw_affiliation_strings":["Intel Corp., Santa Clara, CA, USA"],"raw_orcid":"https://orcid.org/0000-0003-3339-7705","affiliations":[{"raw_affiliation_string":"Intel Corp., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090801235","display_name":"Andrew Boutros","orcid":"https://orcid.org/0000-0002-8044-1644"},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Andrew Boutros","raw_affiliation_strings":["Intel Corp., Santa Clara, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-8044-1644","affiliations":[{"raw_affiliation_string":"Intel Corp., Santa Clara, CA, USA","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5084078152","display_name":"Eriko Nurvitadhi","orcid":"https://orcid.org/0000-0002-2347-9590"},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Eriko Nurvitadhi","raw_affiliation_strings":["Accelerator Architecture Lab, Intel, Hillsboro, OR, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Accelerator Architecture Lab, Intel, Hillsboro, OR, USA","institution_ids":["https://openalex.org/I1343180700"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":2645,"currency":"USD","value_usd":2645},"apc_paid":null,"fwci":3.8262,"has_fulltext":false,"cited_by_count":24,"citation_normalized_percentile":{"value":0.93556538,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":99},"biblio":{"volume":"21","issue":"2","first_page":"49","last_page":"52"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9988999962806702,"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":0.9988999962806702,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9965000152587891,"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/T10500","display_name":"Sparse and Compressive Sensing Techniques","score":0.989799976348877,"subfield":{"id":"https://openalex.org/subfields/2206","display_name":"Computational Mechanics"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8568981885910034},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.7341001629829407},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.7216974496841431},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.7091324329376221},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.6415947675704956},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.4949457049369812},{"id":"https://openalex.org/keywords/node","display_name":"Node (physics)","score":0.4905330538749695},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.47922736406326294},{"id":"https://openalex.org/keywords/bandwidth","display_name":"Bandwidth (computing)","score":0.44687938690185547},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.2481841742992401},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.20283105969429016}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8568981885910034},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7341001629829407},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.7216974496841431},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.7091324329376221},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.6415947675704956},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.4949457049369812},{"id":"https://openalex.org/C62611344","wikidata":"https://www.wikidata.org/wiki/Q1062658","display_name":"Node (physics)","level":2,"score":0.4905330538749695},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.47922736406326294},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.44687938690185547},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.2481841742992401},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.20283105969429016},{"id":"https://openalex.org/C66938386","wikidata":"https://www.wikidata.org/wiki/Q633538","display_name":"Structural 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}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/lca.2022.3189207","is_oa":false,"landing_page_url":"https://doi.org/10.1109/lca.2022.3189207","pdf_url":null,"source":{"id":"https://openalex.org/S17643076","display_name":"IEEE Computer Architecture Letters","issn_l":"1556-6056","issn":["1556-6056","1556-6064","2473-2575"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Computer Architecture Letters","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":13,"referenced_works":["https://openalex.org/W2057332538","https://openalex.org/W2131613942","https://openalex.org/W3025813225","https://openalex.org/W3106981120","https://openalex.org/W3131607724","https://openalex.org/W3166559827","https://openalex.org/W3210457773","https://openalex.org/W4224308101","https://openalex.org/W6750782330","https://openalex.org/W6751609549","https://openalex.org/W6763300512","https://openalex.org/W6785530804","https://openalex.org/W6810081322"],"related_works":["https://openalex.org/W2595172197","https://openalex.org/W2084856301","https://openalex.org/W2127970246","https://openalex.org/W4382618745","https://openalex.org/W2885125400","https://openalex.org/W1001352512","https://openalex.org/W1989889224","https://openalex.org/W1973775000","https://openalex.org/W2011430815","https://openalex.org/W4321606653"],"abstract_inverted_index":{"Training":[0],"state-of-the-art":[1],"artificial":[2],"intelligence":[3],"(AI)":[4],"models":[5],"requires":[6],"scaling":[7],"to":[8,22,73,94],"many":[9],"compute":[10,92],"nodes":[11,42],"and":[12,77,101,116],"relies":[13],"heavily":[14],"on":[15],"collective":[16],"communication":[17,106],"operations,":[18],"such":[19],"as":[20,38],"all-reduce,":[21],"exchange":[23],"the":[24,39,50,90,96,103],"weight":[25],"gradients":[26],"between":[27],"nodes.":[28,139],"The":[29,84],"overhead":[30],"of":[31,41],"these":[32],"operations":[33,76,100],"can":[34],"bottleneck":[35],"training":[36,67,114,127],"performance":[37,128,135],"number":[40],"increases.":[43],"In":[44],"this":[45],"paper,":[46],"we":[47,55],"first":[48],"characterize":[49],"all-reduce":[51,75],"operation":[52],"overhead.":[53],"Then,":[54],"propose":[56],"a":[57,110],"new":[58],"smart":[59,86,123],"network":[60],"interface":[61],"card":[62],"(NIC)":[63],"for":[64],"distributed":[65],"AI":[66,85,113,122],"using":[68],"field-programmable":[69],"gate":[70],"arrays":[71],"(FPGAs)":[72],"accelerate":[74],"optimize":[78],"bandwidth":[79],"utilization":[80],"via":[81],"data":[82],"compression.":[83],"NIC":[87,124],"frees":[88],"up":[89],"system's":[91],"resources":[93],"perform":[95],"more":[97],"compute-intensive":[98],"tensor":[99],"increases":[102],"overall":[104,126],"node-to-node":[105],"efficiency.":[107],"We":[108],"build":[109],"prototype":[111],"6-node":[112],"system":[115],"show":[117],"that":[118],"our":[119],"proposed":[120],"FPGA-based":[121],"enhances":[125],"by":[129],"1.6\u00d7,":[130],"with":[131],"an":[132],"estimated":[133],"2.5\u00d7":[134],"improvement":[136],"at":[137],"32":[138]},"counts_by_year":[{"year":2026,"cited_by_count":6},{"year":2025,"cited_by_count":8},{"year":2024,"cited_by_count":5},{"year":2023,"cited_by_count":3},{"year":2022,"cited_by_count":2}],"updated_date":"2026-08-21T09:56:20.448147","created_date":"2025-10-10T00:00:00"}