{"id":"https://openalex.org/W4312569226","doi":"https://doi.org/10.1109/iscas48785.2022.9937659","title":"An Efficient Hardware Accelerator for Sparse Transformer Neural Networks","display_name":"An Efficient Hardware Accelerator for Sparse Transformer Neural Networks","publication_year":2022,"publication_date":"2022-05-28","ids":{"openalex":"https://openalex.org/W4312569226","doi":"https://doi.org/10.1109/iscas48785.2022.9937659"},"language":"en","primary_location":{"id":"doi:10.1109/iscas48785.2022.9937659","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iscas48785.2022.9937659","pdf_url":null,"source":{"id":"https://openalex.org/S4363604393","display_name":"2022 IEEE International Symposium on Circuits and Systems (ISCAS)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE International Symposium on Circuits and Systems (ISCAS)","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/A5066888122","display_name":"Chao Fang","orcid":"https://orcid.org/0000-0003-3430-1189"},"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":true,"raw_author_name":"Chao Fang","raw_affiliation_strings":["Nanjing University,School of Electronic Science and Engineering,Nanjing,China","School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"affiliations":[{"raw_affiliation_string":"Nanjing University,School of Electronic Science and Engineering,Nanjing,China","institution_ids":["https://openalex.org/I881766915"]},{"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/A5081601767","display_name":"Shouliang Guo","orcid":null},"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":"Shouliang Guo","raw_affiliation_strings":["Nanjing University,School of Electronic Science and Engineering,Nanjing,China","School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"affiliations":[{"raw_affiliation_string":"Nanjing University,School of Electronic Science and Engineering,Nanjing,China","institution_ids":["https://openalex.org/I881766915"]},{"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/A5100715764","display_name":"Wei Wu","orcid":"https://orcid.org/0000-0003-1845-2809"},"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":"Wei Wu","raw_affiliation_strings":["Nanjing University,School of Electronic Science and Engineering,Nanjing,China","School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"affiliations":[{"raw_affiliation_string":"Nanjing University,School of Electronic Science and Engineering,Nanjing,China","institution_ids":["https://openalex.org/I881766915"]},{"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":["Nanjing University,School of Electronic Science and Engineering,Nanjing,China","School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"affiliations":[{"raw_affiliation_string":"Nanjing University,School of Electronic Science and Engineering,Nanjing,China","institution_ids":["https://openalex.org/I881766915"]},{"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/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":["Nanjing University,School of Electronic Science and Engineering,Nanjing,China","School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"affiliations":[{"raw_affiliation_string":"Nanjing University,School of Electronic Science and Engineering,Nanjing,China","institution_ids":["https://openalex.org/I881766915"]},{"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/A5103493548","display_name":"Ming Kai Hsu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ming Kai Hsu","raw_affiliation_strings":["Kuaishou Technology,Palo Alto,CA,USA","Kuaishou Technology, Palo Alto, CA, USA"],"affiliations":[{"raw_affiliation_string":"Kuaishou Technology,Palo Alto,CA,USA","institution_ids":[]},{"raw_affiliation_string":"Kuaishou Technology, Palo Alto, CA, USA","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101421899","display_name":"Lingzhi Liu","orcid":"https://orcid.org/0000-0002-8596-5199"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lingzhi Liu","raw_affiliation_strings":["Kuaishou Technology,Palo Alto,CA,USA","Kuaishou Technology, Palo Alto, CA, USA"],"affiliations":[{"raw_affiliation_string":"Kuaishou Technology,Palo Alto,CA,USA","institution_ids":[]},{"raw_affiliation_string":"Kuaishou Technology, Palo Alto, CA, USA","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5066888122"],"corresponding_institution_ids":["https://openalex.org/I881766915"],"apc_list":null,"apc_paid":null,"fwci":1.4371,"has_fulltext":false,"cited_by_count":26,"citation_normalized_percentile":{"value":0.87941803,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"2670","last_page":"2674"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9984999895095825,"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":0.9984999895095825,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9968000054359436,"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/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.996399998664856,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7818377614021301},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.6997193098068237},{"id":"https://openalex.org/keywords/efficient-energy-use","display_name":"Efficient energy use","score":0.6238536834716797},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.5932859778404236},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5021276473999023},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.48215052485466003},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.4373512864112854},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.41580620408058167},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.3756125271320343},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.370392382144928},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.35262858867645264},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.323772132396698},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.1101427674293518},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.10544276237487793},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.08619582653045654}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7818377614021301},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.6997193098068237},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.6238536834716797},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5932859778404236},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5021276473999023},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.48215052485466003},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.4373512864112854},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.41580620408058167},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.3756125271320343},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.370392382144928},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.35262858867645264},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.323772132396698},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.1101427674293518},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.10544276237487793},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.08619582653045654},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iscas48785.2022.9937659","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iscas48785.2022.9937659","pdf_url":null,"source":{"id":"https://openalex.org/S4363604393","display_name":"2022 IEEE International Symposium on Circuits and Systems (ISCAS)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE International Symposium on Circuits and Systems (ISCAS)","raw_type":"proceedings-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":[],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":17,"referenced_works":["https://openalex.org/W2896457183","https://openalex.org/W2899244816","https://openalex.org/W2964085268","https://openalex.org/W2998135987","https://openalex.org/W3047848469","https://openalex.org/W3094502228","https://openalex.org/W3105966348","https://openalex.org/W3127067080","https://openalex.org/W3159481202","https://openalex.org/W3162542754","https://openalex.org/W3196923642","https://openalex.org/W4385245566","https://openalex.org/W6739901393","https://openalex.org/W6784333009","https://openalex.org/W6789946416","https://openalex.org/W6898505805","https://openalex.org/W6917394425"],"related_works":["https://openalex.org/W1569389315","https://openalex.org/W2503642292","https://openalex.org/W1996757432","https://openalex.org/W2995926156","https://openalex.org/W2063534976","https://openalex.org/W2284838239","https://openalex.org/W2056818248","https://openalex.org/W3094426418","https://openalex.org/W1975010174","https://openalex.org/W2996827035"],"abstract_inverted_index":{"Transformers":[0],"have":[1],"been":[2],"an":[3,34],"indispensable":[4],"staple":[5],"in":[6],"deep":[7],"learning.":[8],"However,":[9],"it":[10,111],"is":[11],"challenging":[12],"to":[13,21,96,115],"realize":[14],"efficient":[15,35],"deployment":[16],"for":[17],"Transformer-based":[18],"model":[19],"due":[20],"their":[22],"substantial":[23],"computation":[24],"and":[25,63,99,117,123],"memory":[26],"demands.":[27],"To":[28],"address":[29],"this":[30],"issue,":[31],"we":[32],"present":[33],"sparse":[36,62],"Transformer":[37],"accelerator":[38],"on":[39,120],"FPGA,":[40],"namely":[41],"STA,":[42],"by":[43],"exploiting":[44],"N:M":[45],"fine-grained":[46],"structured":[47],"sparsity.":[48],"Our":[49],"design":[50],"features":[51],"not":[52],"only":[53],"a":[54,73],"unified":[55],"computing":[56],"engine":[57],"capable":[58],"of":[59],"performing":[60],"both":[61],"dense":[64],"matrix":[65],"multiplications":[66],"with":[67,105],"high":[68],"computational":[69],"efficiency,":[70,125],"but":[71],"also":[72],"scalable":[74],"softmax":[75],"module":[76],"eliminating":[77],"the":[78,92,106],"latency":[79,94],"from":[80],"intermediate":[81],"off-chip":[82],"data":[83],"communication.":[84],"Experimental":[85],"results":[86],"show":[87],"that":[88],"our":[89],"implementation":[90],"achieves":[91],"lowest":[93],"compared":[95,104],"CPU,":[97],"GPU,":[98],"prior":[100],"FPGA-based":[101,109],"accelerators.":[102],"Moreover,":[103],"state-of":[107],"the-art":[108],"accelerators,":[110],"can":[112],"achieve":[113],"up":[114],"$12.28\\times$":[116],"$51.00\\times$":[118],"improvement":[119],"energy":[121],"efficiency":[122],"MAC":[124],"respectively.":[126]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":10},{"year":2024,"cited_by_count":10},{"year":2023,"cited_by_count":3},{"year":2022,"cited_by_count":1}],"updated_date":"2026-03-16T09:10:04.655348","created_date":"2025-10-10T00:00:00"}
