{"id":"https://openalex.org/W3196923642","doi":"https://doi.org/10.1109/socc49529.2020.9524802","title":"Hardware Accelerator for Multi-Head Attention and Position-Wise Feed-Forward in the Transformer","display_name":"Hardware Accelerator for Multi-Head Attention and Position-Wise Feed-Forward in the Transformer","publication_year":2020,"publication_date":"2020-09-08","ids":{"openalex":"https://openalex.org/W3196923642","doi":"https://doi.org/10.1109/socc49529.2020.9524802","mag":"3196923642"},"language":"en","primary_location":{"id":"doi:10.1109/socc49529.2020.9524802","is_oa":false,"landing_page_url":"https://doi.org/10.1109/socc49529.2020.9524802","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 IEEE 33rd International System-on-Chip Conference (SOCC)","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":"https://openalex.org/A5084117570","display_name":"Siyuan Lu","orcid":"https://orcid.org/0000-0002-8639-3081"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Siyuan Lu","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100785858","display_name":"Meiqi Wang","orcid":"https://orcid.org/0000-0001-9553-3640"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Meiqi Wang","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043630267","display_name":"Shuang Liang","orcid":"https://orcid.org/0000-0003-0457-6093"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shuang Liang","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019940239","display_name":"Jun Lin","orcid":"https://orcid.org/0000-0003-2760-4333"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jun Lin","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100696999","display_name":"Zhongfeng Wang","orcid":"https://orcid.org/0000-0002-7227-4786"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhongfeng Wang","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I881766915"],"apc_list":null,"apc_paid":null,"fwci":3.5886,"has_fulltext":false,"cited_by_count":136,"citation_normalized_percentile":{"value":0.94705179,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":99,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"84","last_page":"89"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9991000294685364,"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":0.9991000294685364,"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/T10028","display_name":"Topic Modeling","score":0.9944999814033508,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9901000261306763,"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/computer-science","display_name":"Computer science","score":0.768735408782959},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.7068040370941162},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.6937013268470764},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.6729816198348999},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5916895866394043},{"id":"https://openalex.org/keywords/recurrent-neural-network","display_name":"Recurrent neural network","score":0.5640508532524109},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.5234767198562622},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5053756833076477},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.4215385913848877},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.3532976806163788},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.3384280800819397},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.3247649073600769},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2815830707550049},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.1557578444480896},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.09463846683502197}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.768735408782959},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.7068040370941162},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.6937013268470764},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6729816198348999},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5916895866394043},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.5640508532524109},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.5234767198562622},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5053756833076477},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.4215385913848877},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3532976806163788},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.3384280800819397},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3247649073600769},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2815830707550049},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.1557578444480896},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.09463846683502197},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"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/socc49529.2020.9524802","is_oa":false,"landing_page_url":"https://doi.org/10.1109/socc49529.2020.9524802","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 IEEE 33rd International System-on-Chip Conference (SOCC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.4699999988079071,"display_name":"Industry, innovation and infrastructure","id":"https://metadata.un.org/sdg/9"}],"awards":[{"id":"https://openalex.org/G4988504959","display_name":null,"funder_award_id":"61604068","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":29,"referenced_works":["https://openalex.org/W1924770834","https://openalex.org/W2064675550","https://openalex.org/W2608093348","https://openalex.org/W2896457183","https://openalex.org/W2910396952","https://openalex.org/W2923014074","https://openalex.org/W2938830017","https://openalex.org/W2947946877","https://openalex.org/W2963209464","https://openalex.org/W2963310665","https://openalex.org/W2963341956","https://openalex.org/W2963403868","https://openalex.org/W2966892770","https://openalex.org/W2995923603","https://openalex.org/W2996428491","https://openalex.org/W3017024317","https://openalex.org/W3082274269","https://openalex.org/W3203309275","https://openalex.org/W4288089799","https://openalex.org/W4288337707","https://openalex.org/W4385245566","https://openalex.org/W6640212811","https://openalex.org/W6755207826","https://openalex.org/W6762805369","https://openalex.org/W6766668698","https://openalex.org/W6768021236","https://openalex.org/W6769627184","https://openalex.org/W6780226713","https://openalex.org/W6802232916"],"related_works":["https://openalex.org/W4225394202","https://openalex.org/W4298287631","https://openalex.org/W2518118925","https://openalex.org/W3159273459","https://openalex.org/W4319952061","https://openalex.org/W4280636456","https://openalex.org/W4388913998","https://openalex.org/W4310584535","https://openalex.org/W4295935044","https://openalex.org/W3159906349"],"abstract_inverted_index":{"Designing":[0],"hardware":[1,59,76,128,141,166,181,232],"accelerators":[2,17,233],"for":[3,20,61,78,229,234],"deep":[4],"neural":[5,23,28],"networks":[6,24,29],"(DNNs)":[7],"has":[8,65],"been":[9,67],"much":[10],"desired.":[11],"Nonetheless,":[12],"most":[13,99,125],"of":[14,48,126,143,172,207],"these":[15],"existing":[16],"are":[18,96,159],"built":[19],"either":[21],"convolutional":[22],"(CNNs)":[25],"or":[26],"recurrent":[27],"(RNNs).":[30],"Recently,":[31],"the":[32,37,40,58,62,74,83,89,97,103,113,117,120,127,131,139,144,149,165,170,173,193,198,201,211,218],"Transformer":[33,63,236],"model":[34,64],"is":[35,109,134,148,178],"replacing":[36],"RNN":[38],"in":[39,102,116,152,210,217],"natural":[41],"language":[42,183],"processing":[43],"(NLP)":[44],"area.":[45],"However,":[46],"because":[47],"intensive":[49],"matrix":[50],"computations":[51],"and":[52,88,168,185,214],"complicated":[53,156],"data":[54],"flow":[55,133],"being":[56],"involved,":[57],"design":[60,177,203],"never":[66],"reported.":[68],"In":[69],"this":[70,223],"paper,":[71],"we":[72],"propose":[73],"first":[75],"accelerator":[77],"two":[79,98,121],"key":[80],"components,":[81],"i.e.,":[82],"multi-head":[84],"attention":[85],"(MHA)":[86],"ResBlock":[87],"position-wise":[90],"feed-forward":[91],"network":[92],"(FFN)":[93],"ResBlock,":[94,213,220],"which":[95,147],"complex":[100],"layers":[101],"Transformer.":[104],"Firstly,":[105],"an":[106],"efficient":[107,231],"method":[108],"introduced":[110],"to":[111,123,137,162],"partition":[112],"huge":[114],"matrices":[115],"Transformer,":[118],"allowing":[119],"ResBlocks":[122],"share":[124],"resources.":[129],"Secondly,":[130],"computation":[132],"well":[135],"designed":[136],"ensure":[138],"high":[140],"utilization":[142],"systolic":[145],"array,":[146],"biggest":[150],"module":[151],"our":[153],"design.":[154],"Thirdly,":[155],"nonlinear":[157],"functions":[158],"highly":[160],"optimized":[161],"further":[163],"reduce":[164],"complexity":[167],"also":[169],"latency":[171],"entire":[174],"system.":[175],"Our":[176],"coded":[179],"using":[180],"description":[182],"(HDL)":[184],"evaluated":[186],"on":[187,195],"a":[188,205,226],"Xilinx":[189],"FPGA.":[190],"Compared":[191],"with":[192,197],"implementation":[194],"GPU":[196],"same":[199],"setting,":[200],"proposed":[202],"demonstrates":[204],"speed-up":[206],"14.6":[208],"x":[209,216],"MHA":[212],"3.4":[215],"FFN":[219],"respectively.":[221],"Therefore,":[222],"work":[224],"lays":[225],"good":[227],"foundation":[228],"building":[230],"multiple":[235],"networks.":[237]},"counts_by_year":[{"year":2026,"cited_by_count":10},{"year":2025,"cited_by_count":49},{"year":2024,"cited_by_count":35},{"year":2023,"cited_by_count":26},{"year":2022,"cited_by_count":16}],"updated_date":"2026-08-11T07:18:39.950985","created_date":"2025-10-10T00:00:00"}