{"id":"https://openalex.org/W3193784739","doi":"https://doi.org/10.1109/jssc.2021.3103603","title":"A Neural Network Training Processor With 8-Bit Shared Exponent Bias Floating Point and Multiple-Way Fused Multiply-Add Trees","display_name":"A Neural Network Training Processor With 8-Bit Shared Exponent Bias Floating Point and Multiple-Way Fused Multiply-Add Trees","publication_year":2021,"publication_date":"2021-08-17","ids":{"openalex":"https://openalex.org/W3193784739","doi":"https://doi.org/10.1109/jssc.2021.3103603","mag":"3193784739"},"language":"en","primary_location":{"id":"doi:10.1109/jssc.2021.3103603","is_oa":false,"landing_page_url":"https://doi.org/10.1109/jssc.2021.3103603","pdf_url":null,"source":{"id":"https://openalex.org/S83637746","display_name":"IEEE Journal of Solid-State Circuits","issn_l":"0018-9200","issn":["0018-9200","1558-173X"],"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 Journal of Solid-State Circuits","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/A5089174485","display_name":"Jeongwoo Park","orcid":"https://orcid.org/0000-0002-9603-9588"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jeongwoo Park","raw_affiliation_strings":["Graduate School of Convergence Science and Technology, Seoul National University, Seoul, South Korea","Inter-University Semiconductor Research Center, Seoul National University, Seoul, South Korea"],"raw_orcid":"https://orcid.org/0000-0002-9603-9588","affiliations":[{"raw_affiliation_string":"Graduate School of Convergence Science and Technology, Seoul National University, Seoul, South Korea","institution_ids":[]},{"raw_affiliation_string":"Inter-University Semiconductor Research Center, Seoul National University, Seoul, South Korea","institution_ids":["https://openalex.org/I139264467"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100698555","display_name":"Sunwoo Lee","orcid":"https://orcid.org/0000-0001-7760-0168"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Sunwoo Lee","raw_affiliation_strings":["Graduate School of Convergence Science and Technology, Seoul National University, Seoul, South Korea","Inter-University Semiconductor Research Center, Seoul National University, Seoul, South Korea"],"raw_orcid":"https://orcid.org/0000-0001-7760-0168","affiliations":[{"raw_affiliation_string":"Graduate School of Convergence Science and Technology, Seoul National University, Seoul, South Korea","institution_ids":[]},{"raw_affiliation_string":"Inter-University Semiconductor Research Center, Seoul National University, Seoul, South Korea","institution_ids":["https://openalex.org/I139264467"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5053248335","display_name":"Dongsuk Jeon","orcid":"https://orcid.org/0000-0002-0395-8076"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Dongsuk Jeon","raw_affiliation_strings":["Graduate School of Convergence Science and Technology, Seoul National University, Seoul, South Korea","Inter-University Semiconductor Research Center, Seoul National University, Seoul, South Korea"],"raw_orcid":"https://orcid.org/0000-0002-0395-8076","affiliations":[{"raw_affiliation_string":"Graduate School of Convergence Science and Technology, Seoul National University, Seoul, South Korea","institution_ids":[]},{"raw_affiliation_string":"Inter-University Semiconductor Research Center, Seoul National University, Seoul, South Korea","institution_ids":["https://openalex.org/I139264467"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I139264467"],"apc_list":null,"apc_paid":null,"fwci":1.8844,"has_fulltext":false,"cited_by_count":25,"citation_normalized_percentile":{"value":0.8786299,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":"57","issue":"3","first_page":"965","last_page":"977"},"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/T10320","display_name":"Neural Networks and Applications","score":0.9969000220298767,"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/T12549","display_name":"Image and Object Detection Techniques","score":0.989300012588501,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/exponent","display_name":"Exponent","score":0.7362220883369446},{"id":"https://openalex.org/keywords/bit","display_name":"Bit (key)","score":0.7132071256637573},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6507861614227295},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.6117994785308838},{"id":"https://openalex.org/keywords/point","display_name":"Point (geometry)","score":0.6000185608863831},{"id":"https://openalex.org/keywords/floating-point","display_name":"Floating point","score":0.48177215456962585},{"id":"https://openalex.org/keywords/arithmetic","display_name":"Arithmetic","score":0.448292076587677},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.4437602460384369},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.41180163621902466},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.39624133706092834},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.23138433694839478},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.13993757963180542},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.07681828737258911},{"id":"https://openalex.org/keywords/geometry","display_name":"Geometry","score":0.053384989500045776}],"concepts":[{"id":"https://openalex.org/C2780388253","wikidata":"https://www.wikidata.org/wiki/Q5421508","display_name":"Exponent","level":2,"score":0.7362220883369446},{"id":"https://openalex.org/C117011727","wikidata":"https://www.wikidata.org/wiki/Q1278488","display_name":"Bit (key)","level":2,"score":0.7132071256637573},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6507861614227295},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.6117994785308838},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.6000185608863831},{"id":"https://openalex.org/C84211073","wikidata":"https://www.wikidata.org/wiki/Q117879","display_name":"Floating point","level":2,"score":0.48177215456962585},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.448292076587677},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.4437602460384369},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.41180163621902466},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.39624133706092834},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.23138433694839478},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.13993757963180542},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.07681828737258911},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.053384989500045776},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/jssc.2021.3103603","is_oa":false,"landing_page_url":"https://doi.org/10.1109/jssc.2021.3103603","pdf_url":null,"source":{"id":"https://openalex.org/S83637746","display_name":"IEEE Journal of Solid-State Circuits","issn_l":"0018-9200","issn":["0018-9200","1558-173X"],"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 Journal of Solid-State Circuits","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7","score":0.8999999761581421}],"awards":[{"id":"https://openalex.org/G2351560784","display_name":null,"funder_award_id":"NRF-2019R1C1C1004927","funder_id":"https://openalex.org/F4320322120","funder_display_name":"National Research Foundation of Korea"}],"funders":[{"id":"https://openalex.org/F4320322120","display_name":"National Research Foundation of Korea","ror":"https://ror.org/013aysd81"},{"id":"https://openalex.org/F4320322202","display_name":"IC Design Education Center","ror":"https://ror.org/005v57z85"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":32,"referenced_works":["https://openalex.org/W1999085092","https://openalex.org/W2064675550","https://openalex.org/W2130084210","https://openalex.org/W2194775991","https://openalex.org/W2289252105","https://openalex.org/W2469490737","https://openalex.org/W2535838896","https://openalex.org/W2750822049","https://openalex.org/W2763421725","https://openalex.org/W2898985762","https://openalex.org/W2920954974","https://openalex.org/W2921918777","https://openalex.org/W2921967969","https://openalex.org/W2963122961","https://openalex.org/W2963684088","https://openalex.org/W3015729306","https://openalex.org/W3037742886","https://openalex.org/W3048705738","https://openalex.org/W3048842964","https://openalex.org/W3135859967","https://openalex.org/W3172967059","https://openalex.org/W4297687186","https://openalex.org/W4320930577","https://openalex.org/W6638783484","https://openalex.org/W6685352114","https://openalex.org/W6720242923","https://openalex.org/W6745245109","https://openalex.org/W6746023985","https://openalex.org/W6753770798","https://openalex.org/W6767032739","https://openalex.org/W6780302432","https://openalex.org/W6790866647"],"related_works":["https://openalex.org/W2411923897","https://openalex.org/W4394546135","https://openalex.org/W4255129386","https://openalex.org/W4245068234","https://openalex.org/W2025161972","https://openalex.org/W1978352992","https://openalex.org/W1988006935","https://openalex.org/W1585685783","https://openalex.org/W2119414612","https://openalex.org/W2268486090"],"abstract_inverted_index":{"Recent":[0],"advances":[1],"in":[2,54,143],"deep":[3],"neural":[4,55,59,86,93,220],"networks":[5,56,182],"(DNNs)":[6],"and":[7,29,152,200,224],"machine":[8,19],"learning":[9,20],"algorithms":[10,21],"have":[11,42,51,70],"induced":[12],"the":[13],"demand":[14],"for":[15,35,57,90,102],"services":[16],"based":[17],"on":[18],"that":[22,172,179],"require":[23],"a":[24,74,117,169],"large":[25],"number":[26],"of":[27,180],"computations,":[28],"specialized":[30],"hardware":[31,145],"ranging":[32],"from":[33],"accelerators":[34],"data":[36],"centers":[37],"to":[38,72,111,147,167,175,205],"on-device":[39],"computing":[40],"systems":[41],"been":[43,52],"introduced.":[44,134],"Low-precision":[45],"math":[46,199],"such":[47],"as":[48,124],"8-bit":[49,84,97,125],"integers":[50],"used":[53,142],"energy-efficient":[58],"network":[60,87,221],"inference,":[61],"but":[62],"training":[63,88,103,190,222],"with":[64,128,158,184,197],"low-precision":[65],"numbers":[66,99],"without":[67],"performance":[68,171],"degradation":[69],"remained":[71],"be":[73],"challenge.":[75],"To":[76],"overcome":[77],"this":[78,80],"challenge,":[79],"article":[81],"presents":[82],"an":[83],"floating-point":[85,98],"processor":[89,191],"state-of-the-art":[91],"non-sparse":[92],"networks.":[94],"As":[95],"na\u00efve":[96],"are":[100,109,141],"insufficient":[101],"DNNs":[104],"robustly,":[105],"two":[106],"additional":[107],"methods":[108],"introduced":[110],"ensure":[112,148],"high-performance":[113],"DNN":[114,189],"training.":[115],"First,":[116],"novel":[118],"numeric":[119],"system":[120],"which":[121],"we":[122],"dub":[123],"floating":[126],"point":[127],"shared":[129],"exponent":[130],"bias":[131],"(FP8-SEB)":[132],"is":[133,162,173],"Moreover,":[135],"multiple-way":[136,159],"fused":[137],"multiply-add":[138],"(FMA)":[139],"trees":[140,161,195],"FP8-SEB\u2019s":[144],"implementation":[146],"higher":[149,214],"numerical":[150],"precision":[151],"reduced":[153],"energy.":[154],"FP8-SEB":[155,198],"format":[156],"combined":[157],"FMA":[160,194],"evaluated":[163],"under":[164],"various":[165],"scenarios":[166],"show":[168,206],"trained-from-scratch":[170],"close":[174],"or":[176],"even":[177],"surpasses":[178],"current":[181],"trained":[183],"full-precision":[185],"(FP32).":[186],"Our":[187],"silicon-verified":[188],"utilizes":[192],"24-way":[193],"implemented":[196],"flexible":[201],"2-D":[202],"routing":[203],"schemes":[204],"<inline-formula":[207,225],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[208,226],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">":[209,227],"<tex-math":[210,228],"notation=\"LaTeX\">$2.48\\times":[211],"$":[212,230],"</tex-math></inline-formula>":[213,231],"energy":[215,233],"efficiency":[216],"than":[217,234],"prior":[218],"low-power":[219],"processors":[223],"notation=\"LaTeX\">$78.1\\times":[229],"lower":[232],"standard":[235],"GPUs.":[236]},"counts_by_year":[{"year":2025,"cited_by_count":5},{"year":2024,"cited_by_count":6},{"year":2023,"cited_by_count":7},{"year":2022,"cited_by_count":7}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}