{"id":"https://openalex.org/W4413755340","doi":"https://doi.org/10.1109/isvlsi65124.2025.11130197","title":"PACE: An Optimal Piecewise Polynomial Approximation Unit for Flexible and Efficient Transformer Non-linearity Acceleration","display_name":"PACE: An Optimal Piecewise Polynomial Approximation Unit for Flexible and Efficient Transformer Non-linearity Acceleration","publication_year":2025,"publication_date":"2025-07-06","ids":{"openalex":"https://openalex.org/W4413755340","doi":"https://doi.org/10.1109/isvlsi65124.2025.11130197"},"language":"en","primary_location":{"id":"doi:10.1109/isvlsi65124.2025.11130197","is_oa":false,"landing_page_url":"https://doi.org/10.1109/isvlsi65124.2025.11130197","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","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/A5056197249","display_name":"Arpan Suravi Prasad","orcid":"https://orcid.org/0009-0009-6031-6668"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":true,"raw_author_name":"Arpan Suravi Prasad","raw_affiliation_strings":["ETH Zurich,IIS,Switzerland"],"affiliations":[{"raw_affiliation_string":"ETH Zurich,IIS,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073926165","display_name":"Gamze \u0130slamo\u011flu","orcid":"https://orcid.org/0000-0002-5129-1691"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Gamze \u0130slamo\u011flu","raw_affiliation_strings":["ETH Zurich,IIS,Switzerland"],"affiliations":[{"raw_affiliation_string":"ETH Zurich,IIS,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5062088050","display_name":"Luca Bertaccini","orcid":"https://orcid.org/0000-0002-3011-6368"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Luca Bertaccini","raw_affiliation_strings":["ETH Zurich,IIS,Switzerland"],"affiliations":[{"raw_affiliation_string":"ETH Zurich,IIS,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023905268","display_name":"Davide Rossi","orcid":"https://orcid.org/0000-0002-0651-5393"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Davide Rossi","raw_affiliation_strings":["ETH Zurich,IIS,Switzerland"],"affiliations":[{"raw_affiliation_string":"ETH Zurich,IIS,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5089543048","display_name":"Francesco Conti","orcid":null},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Francesco Conti","raw_affiliation_strings":["ETH Zurich,IIS,Switzerland"],"affiliations":[{"raw_affiliation_string":"ETH Zurich,IIS,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5043408422","display_name":"Luca Benini","orcid":"https://orcid.org/0000-0001-8068-3806"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Luca Benini","raw_affiliation_strings":["ETH Zurich,IIS,Switzerland"],"affiliations":[{"raw_affiliation_string":"ETH Zurich,IIS,Switzerland","institution_ids":["https://openalex.org/I35440088"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5056197249"],"corresponding_institution_ids":["https://openalex.org/I35440088"],"apc_list":null,"apc_paid":null,"fwci":0.699,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.74121604,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":95},"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/T10323","display_name":"Analog and Mixed-Signal Circuit Design","score":0.979200005531311,"subfield":{"id":"https://openalex.org/subfields/2204","display_name":"Biomedical Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10323","display_name":"Analog and Mixed-Signal Circuit Design","score":0.979200005531311,"subfield":{"id":"https://openalex.org/subfields/2204","display_name":"Biomedical Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12300","display_name":"Advanced Electrical Measurement Techniques","score":0.9606999754905701,"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"}},{"id":"https://openalex.org/T11034","display_name":"Digital Filter Design and Implementation","score":0.9562000036239624,"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/piecewise","display_name":"Piecewise","score":0.6929376721382141},{"id":"https://openalex.org/keywords/linearity","display_name":"Linearity","score":0.6567399501800537},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.6356245279312134},{"id":"https://openalex.org/keywords/polynomial","display_name":"Polynomial","score":0.6040173768997192},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.48771652579307556},{"id":"https://openalex.org/keywords/control-theory","display_name":"Control theory (sociology)","score":0.43858253955841064},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.4022020101547241},{"id":"https://openalex.org/keywords/mathematical-optimization","display_name":"Mathematical optimization","score":0.3804307281970978},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.3534803092479706},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.22229808568954468},{"id":"https://openalex.org/keywords/mathematical-analysis","display_name":"Mathematical analysis","score":0.20278847217559814},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.17939868569374084},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.16128131747245789},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.1301182210445404},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.07818752527236938}],"concepts":[{"id":"https://openalex.org/C164660894","wikidata":"https://www.wikidata.org/wiki/Q2037833","display_name":"Piecewise","level":2,"score":0.6929376721382141},{"id":"https://openalex.org/C77170095","wikidata":"https://www.wikidata.org/wiki/Q1753188","display_name":"Linearity","level":2,"score":0.6567399501800537},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.6356245279312134},{"id":"https://openalex.org/C90119067","wikidata":"https://www.wikidata.org/wiki/Q43260","display_name":"Polynomial","level":2,"score":0.6040173768997192},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.48771652579307556},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.43858253955841064},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4022020101547241},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3804307281970978},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3534803092479706},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.22229808568954468},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.20278847217559814},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.17939868569374084},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.16128131747245789},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.1301182210445404},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.07818752527236938},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.0},{"id":"https://openalex.org/C74650414","wikidata":"https://www.wikidata.org/wiki/Q11397","display_name":"Classical mechanics","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/isvlsi65124.2025.11130197","is_oa":false,"landing_page_url":"https://doi.org/10.1109/isvlsi65124.2025.11130197","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","raw_type":"proceedings-article"},{"id":"pmh:oai:cris.unibo.it:11585/1040835","is_oa":false,"landing_page_url":"https://hdl.handle.net/11585/1040835","pdf_url":null,"source":{"id":"https://openalex.org/S4306402579","display_name":"Archivio istituzionale della ricerca (Alma Mater Studiorum Universit\u00e0 di Bologna)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I4210117483","host_organization_name":"Istituto di Ematologia di Bologna","host_organization_lineage":["https://openalex.org/I4210117483"],"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":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Affordable and clean energy","score":0.4399999976158142,"id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4308757930","https://openalex.org/W2383844134","https://openalex.org/W2374198922","https://openalex.org/W2889599571","https://openalex.org/W2151524381","https://openalex.org/W1833822726","https://openalex.org/W2028757439","https://openalex.org/W2798137860","https://openalex.org/W2065958989","https://openalex.org/W1998603635"],"abstract_inverted_index":{"Diverse":[0],"Transformer":[1],"models":[2],"are":[3,40],"being":[4],"explored":[5],"for":[6],"Contextual":[7],"and":[8,13,24,37,59,81,100,161],"Generative":[9],"AI.":[10],"While":[11],"matrix-matrix":[12,36],"matrixvector":[14,38],"multiplication":[15,39],"remain":[16],"invariant":[17],"core":[18,136],"operators,":[19,51],"non-linear":[20,50,70],"activation":[21],"functions":[22,71],"change":[23],"evolve":[25],"very":[26],"rapidly,":[27],"emerging":[28],"as":[29,35,52,108],"a":[30,113,152],"new":[31],"bottleneck":[32],"-":[33],"especially":[34],"increasingly":[41],"accelerated.":[42],"This":[43],"necessitates":[44],"hardware":[45],"flexibility":[46],"to":[47,129,176],"accelerate":[48],"diverse":[49],"software":[53],"emulation":[54],"is":[55],"not":[56],"sufficiently":[57],"fast":[58],"efficient.":[60],"We":[61],"present":[62],"PACE\u2014an":[63],"open-source":[64],"Polynomial":[65,75],"Approximation":[66,76],"Compute":[67],"Engine\u2014that":[68],"accelerates":[69],"using":[72],"optimal":[73],"Piecewise":[74],"(PwPA)":[77],"with":[78,140],"arbitrary":[79],"degrees":[80],"partitions,":[82],"achieving":[83],"$\\lt":[84],"0.2":[85],"\\%$":[86,120],"accuracy":[87],"loss":[88],"on":[89],"state-of-the-art":[90],"(SoA)":[91],"pretrained":[92],"Convolutional":[93],"Neural":[94],"Networks":[95],"(CNNs),":[96],"Vision":[97],"Transformers":[98],"(ViTs),":[99],"Large":[101],"Language":[102],"Models":[103],"(LLMs)":[104],"without":[105],"fine-tuning.":[106],"Integrated":[107],"an":[109,141],"ISA":[110],"extension":[111],"into":[112],"RISCV":[114],"processing":[115],"cluster,":[116],"PACE":[117,167],"incurs":[118],"$14":[119],"area":[121],"overhead":[122],"of":[123,143],"the":[124,157,170],"cluster":[125,150],"while":[126],"delivering":[127],"up":[128],"1":[130],"FP32":[131],"PolyEval":[132],"per":[133,135],"cycle":[134],"at":[137],"0.95":[138],"GHz":[139],"efficiency":[142],"$8":[144],"\\mathrm{pJ}":[145],"/$":[146],"PolyEval.":[147],"The":[148],"PACEaugmented":[149],"achieves":[151],"$750":[153],"\\times$":[154,163],"speedup":[155],"over":[156,164],"C":[158],"math.h":[159],"library":[160],"$65":[162],"software-based":[165],"PwPA.":[166],"also":[168],"delivers":[169],"highest":[171],"throughput":[172],"($1.5":[173],"\\times$)":[174],"compared":[175],"existing":[177],"generalpurpose":[178],"approximation":[179],"hardware.":[180]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}
