{"id":"https://openalex.org/W4411725757","doi":"https://doi.org/10.1109/iscas56072.2025.11043187","title":"Efficient Multiple-Precision Floating-Point Multiply-Add Architecture for Deep Learning Applications","display_name":"Efficient Multiple-Precision Floating-Point Multiply-Add Architecture for Deep Learning Applications","publication_year":2025,"publication_date":"2025-05-25","ids":{"openalex":"https://openalex.org/W4411725757","doi":"https://doi.org/10.1109/iscas56072.2025.11043187"},"language":"en","primary_location":{"id":"doi:10.1109/iscas56072.2025.11043187","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iscas56072.2025.11043187","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Symposium on Circuits and Systems (ISCAS)","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":null,"display_name":"Songtai Liang","orcid":null},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Songtai Liang","raw_affiliation_strings":["Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004097849","display_name":"Bingjie Xia","orcid":null},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bingjie Xia","raw_affiliation_strings":["Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5115596220","display_name":"Wen Wang","orcid":"https://orcid.org/0000-0002-2350-3562"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wen Wang","raw_affiliation_strings":["Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101856331","display_name":"Xiaoxia Han","orcid":"https://orcid.org/0009-0001-6832-8948"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiaoxia Han","raw_affiliation_strings":["Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5011627918","display_name":"Peng Liu","orcid":"https://orcid.org/0000-0002-8774-7398"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Peng Liu","raw_affiliation_strings":["Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University,College of Information Science and Electronic Engineering,Hangzhou,China,310027","institution_ids":["https://openalex.org/I76130692"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I76130692"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11697","display_name":"Numerical Methods and Algorithms","score":0.9760000109672546,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T11697","display_name":"Numerical Methods and Algorithms","score":0.9760000109672546,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T11034","display_name":"Digital Filter Design and Implementation","score":0.9726999998092651,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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":0.6917808651924133},{"id":"https://openalex.org/keywords/architecture","display_name":"Architecture","score":0.626073956489563},{"id":"https://openalex.org/keywords/floating-point","display_name":"Floating point","score":0.5207183957099915},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.5179259181022644},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.49974894523620605},{"id":"https://openalex.org/keywords/point","display_name":"Point (geometry)","score":0.4980504512786865},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4405378997325897},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.21776819229125977},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.10102713108062744}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6917808651924133},{"id":"https://openalex.org/C123657996","wikidata":"https://www.wikidata.org/wiki/Q12271","display_name":"Architecture","level":2,"score":0.626073956489563},{"id":"https://openalex.org/C84211073","wikidata":"https://www.wikidata.org/wiki/Q117879","display_name":"Floating point","level":2,"score":0.5207183957099915},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.5179259181022644},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.49974894523620605},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.4980504512786865},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4405378997325897},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.21776819229125977},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.10102713108062744},{"id":"https://openalex.org/C142362112","wikidata":"https://www.wikidata.org/wiki/Q735","display_name":"Art","level":0,"score":0.0},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0},{"id":"https://openalex.org/C153349607","wikidata":"https://www.wikidata.org/wiki/Q36649","display_name":"Visual arts","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iscas56072.2025.11043187","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iscas56072.2025.11043187","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Symposium on Circuits and Systems (ISCAS)","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":13,"referenced_works":["https://openalex.org/W2151215538","https://openalex.org/W2194775991","https://openalex.org/W2604319603","https://openalex.org/W2606722458","https://openalex.org/W2914973130","https://openalex.org/W3217237164","https://openalex.org/W4297097366","https://openalex.org/W4312872667","https://openalex.org/W4386825187","https://openalex.org/W4400233430","https://openalex.org/W6745245109","https://openalex.org/W6763653508","https://openalex.org/W6801071181"],"related_works":["https://openalex.org/W4375867731","https://openalex.org/W2611989081","https://openalex.org/W2731899572","https://openalex.org/W4230611425","https://openalex.org/W4294635752","https://openalex.org/W4304166257","https://openalex.org/W4383066092","https://openalex.org/W3215138031","https://openalex.org/W2588198209","https://openalex.org/W2038503502"],"abstract_inverted_index":{"To":[0],"fully":[1],"exploit":[2],"the":[3,67,99,112,137,142,153],"potential":[4],"of":[5,70,105,161],"parallel":[6,68],"computing":[7],"while":[8,140],"minimizing":[9],"hardware":[10,34],"costs,":[11],"floating-point":[12,46],"multiply-add":[13,86],"(FMA)":[14],"units":[15],"that":[16,49],"support":[17],"multiple":[18],"precisions":[19],"are":[20],"widely":[21],"used":[22],"in":[23],"deep":[24,162],"learning.":[25],"However,":[26],"achieving":[27],"a":[28,37,43,158],"balance":[29],"between":[30],"accuracy,":[31],"performance,":[32],"and":[33,60,94,109],"overhead":[35],"remains":[36],"significant":[38],"challenge.":[39],"This":[40],"paper":[41],"presents":[42],"unified":[44],"multiple-precision":[45,128],"FMA":[47,73,77,81,129],"architecture":[48,64,100,132,155],"supports":[50],"four":[51,113],"precision":[52,114],"formats:":[53],"single-precision":[54],"(SP),":[55],"Bfloat16":[56],"(BF16),":[57],"TensorFloat32+":[58],"(TF32+),":[59],"INT8.":[61],"The":[62,147],"proposed":[63,154],"allows":[65],"for":[66,111,157],"execution":[69],"nine":[71,84,135],"BF16":[72,138],"operations,":[74,78],"two":[75],"TF32+":[76],"one":[79],"SP":[80],"operation,":[82],"or":[83],"INT8":[85],"(MA)":[87],"operations.":[88],"Through":[89],"careful":[90],"data":[91,149],"format":[92,150],"selection":[93],"an":[95],"optimized":[96],"architectural":[97],"design,":[98],"achieves":[101],"multiplier":[102],"utilization":[103],"rates":[104],"100%,":[106,108],"88.9%,":[107],"100%":[110],"modes,":[115],"respectively,":[116],"with":[117],"all":[118],"multipliers":[119],"operating":[120],"at":[121],"full":[122],"bit":[123],"width.":[124],"Compared":[125],"to":[126],"state-of-the-art":[127],"designs,":[130],"this":[131],"delivers":[133],"over":[134],"times":[136],"throughput":[139],"increasing":[141],"area":[143],"by":[144],"only":[145],"49%.":[146],"flexible":[148],"configuration":[151],"makes":[152],"suitable":[156],"wide":[159],"range":[160],"learning":[163],"applications.":[164]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}