{"id":"https://openalex.org/W4386764151","doi":"https://doi.org/10.1109/dac56929.2023.10247855","title":"Flex-SFU: Accelerating DNN Activation Functions by Non-Uniform Piecewise Approximation","display_name":"Flex-SFU: Accelerating DNN Activation Functions by Non-Uniform Piecewise Approximation","publication_year":2023,"publication_date":"2023-07-09","ids":{"openalex":"https://openalex.org/W4386764151","doi":"https://doi.org/10.1109/dac56929.2023.10247855"},"language":"en","primary_location":{"id":"doi:10.1109/dac56929.2023.10247855","doi":"https://doi.org/10.1109/dac56929.2023.10247855","is_oa":false,"landing_page_url":"https://doi.org/10.1109/dac56929.2023.10247855","pdf_url":null,"source":{"id":"https://openalex.org/S4210231368","display_name":"ACM/IEEE Design Automation Conference (DAC)","issn_l":"0738-100X","issn":["0738-100X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core"],"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2023 60th ACM/IEEE Design Automation Conference (DAC)","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2305.04546","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5064165125","display_name":"Enrico Reggiani","orcid":"https://orcid.org/0000-0003-1385-7962","observed_orcids":["https://orcid.org/0000-0003-1385-7962"]},"institutions":[{"id":"https://openalex.org/I2799803557","display_name":"Barcelona Supercomputing Center","ror":"https://ror.org/05sd8tv96","country_code":"ES","type":"facility","lineage":["https://openalex.org/I2799803557","https://openalex.org/I9617848"]}],"countries":["ES"],"is_corresponding":false,"raw_author_name":"Enrico Reggiani","raw_affiliation_strings":["Huawei Zurich Research Center,Computing Systems Lab,Switzerland","Barcelona Supercomputing Center, Spain","Computing Systems Lab, Huawei Zurich Research Center, Switzerland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Zurich Research Center,Computing Systems Lab,Switzerland","institution_ids":[]},{"raw_affiliation_string":"Barcelona Supercomputing Center, Spain","institution_ids":["https://openalex.org/I2799803557"]},{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center, Switzerland","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050957615","display_name":"Renzo Andri","orcid":"https://orcid.org/0000-0002-8776-5158","observed_orcids":["https://orcid.org/0000-0002-8776-5158"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Renzo Andri","raw_affiliation_strings":["Huawei Zurich Research Center,Computing Systems Lab,Switzerland","Computing Systems Lab, Huawei Zurich Research Center, Switzerland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Zurich Research Center,Computing Systems Lab,Switzerland","institution_ids":[]},{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center, Switzerland","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5025399641","display_name":"Lukas Cavigelli","orcid":"https://orcid.org/0000-0003-1767-7715","observed_orcids":["https://orcid.org/0000-0003-1767-7715"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lukas Cavigelli","raw_affiliation_strings":["Huawei Zurich Research Center,Computing Systems Lab,Switzerland","Computing Systems Lab, Huawei Zurich Research Center, Switzerland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Zurich Research Center,Computing Systems Lab,Switzerland","institution_ids":[]},{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center, Switzerland","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.5481,"has_fulltext":true,"cited_by_count":29,"citation_normalized_percentile":{"value":0.93482911,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":93,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.5669000148773193,"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":0.5669000148773193,"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":0.26269999146461487,"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/T10363","display_name":"Low-power high-performance VLSI design","score":0.06599999964237213,"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/activation-function-approximation","display_name":"activation function approximation","score":0.75},{"id":"https://openalex.org/keywords/mean-squared-error","display_name":"mean squared error","score":0.5429999828338623},{"id":"https://openalex.org/keywords/computer-vision-models","display_name":"computer vision models","score":0.6259999871253967},{"id":"https://openalex.org/keywords/natural-language-processing-models","display_name":"natural language processing models","score":0.9950000047683716}],"concepts":[{"id":"https://openalex.org/C137800194","wikidata":"https://www.wikidata.org/wiki/Q11713455","display_name":"Interpolation (computer graphics)","level":3,"score":0.7200688719749451},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6460756063461304},{"id":"https://openalex.org/C164660894","wikidata":"https://www.wikidata.org/wiki/Q2037833","display_name":"Piecewise","level":2,"score":0.6268621683120728},{"id":"https://openalex.org/C84211073","wikidata":"https://www.wikidata.org/wiki/Q117879","display_name":"Floating point","level":2,"score":0.48415452241897583},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4816835820674896},{"id":"https://openalex.org/C122383733","wikidata":"https://www.wikidata.org/wiki/Q865920","display_name":"Approximation error","level":2,"score":0.4574735164642334},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.4476003646850586},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.4471323490142822},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.4420841336250305},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.23930037021636963},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.22390460968017578},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.1758694350719452},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C104114177","wikidata":"https://www.wikidata.org/wiki/Q79782","display_name":"Motion (physics)","level":2,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/dac56929.2023.10247855","doi":"https://doi.org/10.1109/dac56929.2023.10247855","is_oa":false,"landing_page_url":"https://doi.org/10.1109/dac56929.2023.10247855","pdf_url":null,"source":{"id":"https://openalex.org/S4210231368","display_name":"ACM/IEEE Design Automation Conference (DAC)","issn_l":"0738-100X","issn":["0738-100X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core"],"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2023 60th ACM/IEEE Design Automation Conference (DAC)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2305.04546","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2305.04546","pdf_url":"https://arxiv.org/pdf/2305.04546","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2305.04546","doi":"https://doi.org/10.48550/arxiv.2305.04546","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2305.04546","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2305.04546","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2305.04546","pdf_url":"https://arxiv.org/pdf/2305.04546","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4386764151.pdf","grobid_xml":"https://content.openalex.org/works/W4386764151.grobid-xml"},"referenced_works_count":26,"referenced_works":["https://openalex.org/W1522301498","https://openalex.org/W1686810756","https://openalex.org/W1890351280","https://openalex.org/W2000956074","https://openalex.org/W2042519306","https://openalex.org/W2073256350","https://openalex.org/W2120052858","https://openalex.org/W2163605009","https://openalex.org/W2167425979","https://openalex.org/W2169621283","https://openalex.org/W2194775991","https://openalex.org/W2606722458","https://openalex.org/W2896478499","https://openalex.org/W2922301133","https://openalex.org/W2970971581","https://openalex.org/W2982219368","https://openalex.org/W3008360848","https://openalex.org/W3023212902","https://openalex.org/W3036878841","https://openalex.org/W3043675702","https://openalex.org/W3054316493","https://openalex.org/W3160664745","https://openalex.org/W4200126970","https://openalex.org/W4283451911","https://openalex.org/W4283791586","https://openalex.org/W4307079523"],"related_works":["https://openalex.org/W4409154894","https://openalex.org/W7202389582","https://openalex.org/W4291214146","https://openalex.org/W4400880887","https://openalex.org/W3217128056","https://openalex.org/W4417201829","https://openalex.org/W7171766816","https://openalex.org/W4406061168","https://openalex.org/W7171791652","https://openalex.org/W4400234058"],"abstract_inverted_index":{"Modern":[0],"DNN":[1],"workloads":[2],"increasingly":[3],"rely":[4],"on":[5,89,120],"activation":[6,35],"functions":[7,36],"consisting":[8],"of":[9,126,157],"computationally":[10],"complex":[11],"operations.":[12],"This":[13,26],"poses":[14],"a":[15,30,54],"challenge":[16],"to":[17,70,81,97,135,162],"current":[18],"accelerators":[19,130],"optimized":[20],"for":[21,34],"convolutions":[22],"and":[23,47,84,111,149,154,159],"matrix-matrix":[24],"multiplications.":[25],"work":[27],"presents":[28],"Flex-SFU,":[29],"lightweight":[31],"hardware":[32,129],"accelerator":[33],"implementing":[37,53],"non-uniform":[38,82],"piecewise":[39,99],"interpolation":[40,73,83,101],"supporting":[41],"multiple":[42],"data":[43],"formats.":[44],"Non-Uniform":[45],"segments":[46],"floating-point":[48,85],"numbers":[49],"are":[50],"enabled":[51],"by":[52,131],"binary-tree":[55],"comparison":[56],"within":[57],"the":[58,72,76,123,142,163],"address":[59],"decoding":[60],"unit.":[61,167],"An":[62],"SGD-based":[63],"optimization":[64],"algorithm":[65],"with":[66,105,138],"heuristics":[67],"is":[68],"proposed":[69],"find":[71],"function":[74],"reducing":[75],"mean":[77,93],"squared":[78,94],"error.":[79],"Thanks":[80],"support,":[86],"Flex-SFU":[87,118],"achieves":[88],"average":[90],"22.3x":[91],"better":[92],"error":[95],"compared":[96],"previous":[98],"linear":[100],"approaches.":[102],"The":[103],"evaluation":[104],"more":[106],"than":[107],"700":[108],"computer":[109],"vision":[110],"natural":[112],"language":[113],"processing":[114,166],"models":[115],"shows":[116],"that":[117],"can,":[119],"average,":[121],"improve":[122],"end-to-end":[124],"performance":[125],"state-of-the-art":[127],"AI":[128],"35.7%,":[132],"achieving":[133],"up":[134],"3.3x":[136],"speedup":[137],"negligible":[139],"impact":[140],"in":[141],"models\u2019":[143],"accuracy":[144],"when":[145],"using":[146],"32":[147],"segments,":[148],"only":[150],"introducing":[151],"an":[152],"area":[153],"power":[155],"overhead":[156],"5.9%":[158],"0.8%":[160],"relative":[161],"baseline":[164],"vector":[165]},"counts_by_year":[{"year":2026,"cited_by_count":13},{"year":2025,"cited_by_count":10},{"year":2024,"cited_by_count":4},{"year":2023,"cited_by_count":2}],"updated_date":"2026-11-11T12:16:26.678958","created_date":"2023-09-16T00:00:00"}