{"id":"https://openalex.org/W4404955604","doi":"https://doi.org/10.1109/vlsi-soc62099.2024.10767806","title":"Adaptive Block-Scaled GeMMs on Vector Processors for DNN Training at the Edge","display_name":"Adaptive Block-Scaled GeMMs on Vector Processors for DNN Training at the Edge","publication_year":2024,"publication_date":"2024-10-06","ids":{"openalex":"https://openalex.org/W4404955604","doi":"https://doi.org/10.1109/vlsi-soc62099.2024.10767806"},"language":"en","primary_location":{"id":"doi:10.1109/vlsi-soc62099.2024.10767806","is_oa":false,"landing_page_url":"https://doi.org/10.1109/vlsi-soc62099.2024.10767806","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IFIP/IEEE 32nd International Conference on Very Large Scale Integration (VLSI-SoC)","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/A5030379454","display_name":"Nitish Satya Murthy","orcid":"https://orcid.org/0000-0003-0181-8069"},"institutions":[{"id":"https://openalex.org/I4210114974","display_name":"IMEC","ror":"https://ror.org/02kcbn207","country_code":"BE","type":"nonprofit","lineage":["https://openalex.org/I4210114974"]}],"countries":["BE"],"is_corresponding":true,"raw_author_name":"Nitish Satya Murthy","raw_affiliation_strings":["imec,Leuven,Belgium,3001"],"affiliations":[{"raw_affiliation_string":"imec,Leuven,Belgium,3001","institution_ids":["https://openalex.org/I4210114974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5059602336","display_name":"Nathan Laubeuf","orcid":"https://orcid.org/0000-0002-1592-755X"},"institutions":[{"id":"https://openalex.org/I4210114974","display_name":"IMEC","ror":"https://ror.org/02kcbn207","country_code":"BE","type":"nonprofit","lineage":["https://openalex.org/I4210114974"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Nathan Laubeuf","raw_affiliation_strings":["imec,Leuven,Belgium,3001"],"affiliations":[{"raw_affiliation_string":"imec,Leuven,Belgium,3001","institution_ids":["https://openalex.org/I4210114974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074456876","display_name":"Debjyoti Bhattacharjee","orcid":"https://orcid.org/0000-0001-6561-8934"},"institutions":[{"id":"https://openalex.org/I4210114974","display_name":"IMEC","ror":"https://ror.org/02kcbn207","country_code":"BE","type":"nonprofit","lineage":["https://openalex.org/I4210114974"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Debjyoti Bhattacharjee","raw_affiliation_strings":["imec,Leuven,Belgium,3001"],"affiliations":[{"raw_affiliation_string":"imec,Leuven,Belgium,3001","institution_ids":["https://openalex.org/I4210114974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033235745","display_name":"Francky Catthoor","orcid":"https://orcid.org/0000-0002-3599-8515"},"institutions":[{"id":"https://openalex.org/I4210114974","display_name":"IMEC","ror":"https://ror.org/02kcbn207","country_code":"BE","type":"nonprofit","lineage":["https://openalex.org/I4210114974"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Francky Catthoor","raw_affiliation_strings":["imec,Leuven,Belgium,3001"],"affiliations":[{"raw_affiliation_string":"imec,Leuven,Belgium,3001","institution_ids":["https://openalex.org/I4210114974"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5012150553","display_name":"Marian Verhelst","orcid":"https://orcid.org/0000-0003-3495-9263"},"institutions":[{"id":"https://openalex.org/I4210114974","display_name":"IMEC","ror":"https://ror.org/02kcbn207","country_code":"BE","type":"nonprofit","lineage":["https://openalex.org/I4210114974"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Marian Verhelst","raw_affiliation_strings":["imec,Leuven,Belgium,3001"],"affiliations":[{"raw_affiliation_string":"imec,Leuven,Belgium,3001","institution_ids":["https://openalex.org/I4210114974"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5030379454"],"corresponding_institution_ids":["https://openalex.org/I4210114974"],"apc_list":null,"apc_paid":null,"fwci":0.5828,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.70371839,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":99},"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/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9387999773025513,"subfield":{"id":"https://openalex.org/subfields/2808","display_name":"Neurology"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},"topics":[{"id":"https://openalex.org/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9387999773025513,"subfield":{"id":"https://openalex.org/subfields/2808","display_name":"Neurology"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7598814964294434},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.7303876876831055},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.6179688572883606},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.612743079662323},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.4327293038368225},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.38962364196777344},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.12723475694656372}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7598814964294434},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.7303876876831055},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.6179688572883606},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.612743079662323},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4327293038368225},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.38962364196777344},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.12723475694656372},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","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}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/vlsi-soc62099.2024.10767806","is_oa":false,"landing_page_url":"https://doi.org/10.1109/vlsi-soc62099.2024.10767806","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IFIP/IEEE 32nd International Conference on Very Large Scale Integration (VLSI-SoC)","raw_type":"proceedings-article"},{"id":"pmh:oai:lirias2repo.kuleuven.be:20.500.12942/755853","is_oa":false,"landing_page_url":"https://lirias.kuleuven.be/handle/20.500.12942/755853","pdf_url":null,"source":{"id":"https://openalex.org/S4306401954","display_name":"Lirias (KU Leuven)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I99464096","host_organization_name":"KU Leuven","host_organization_lineage":["https://openalex.org/I99464096"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"International Conference on Very Large Scale Integration (VLSI-SoC), Tanger, Morocco, 6-9 October 2024","raw_type":"info:eu-repo/semantics/publishedVersion"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320327336","display_name":"Vlaamse regering","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":22,"referenced_works":["https://openalex.org/W2613264175","https://openalex.org/W2809624076","https://openalex.org/W2967240643","https://openalex.org/W3003537648","https://openalex.org/W3015764133","https://openalex.org/W3176714321","https://openalex.org/W3184606595","https://openalex.org/W3208633927","https://openalex.org/W4294691337","https://openalex.org/W4319301138","https://openalex.org/W4380874652","https://openalex.org/W4386322998","https://openalex.org/W4386764384","https://openalex.org/W4400412147","https://openalex.org/W6745245109","https://openalex.org/W6748522178","https://openalex.org/W6751494529","https://openalex.org/W6761694480","https://openalex.org/W6779824479","https://openalex.org/W6785530804","https://openalex.org/W6790866647","https://openalex.org/W6857288518"],"related_works":["https://openalex.org/W230091440","https://openalex.org/W2233261550","https://openalex.org/W2810751659","https://openalex.org/W258997015","https://openalex.org/W2997094352","https://openalex.org/W3216976533","https://openalex.org/W100620283","https://openalex.org/W2495260952","https://openalex.org/W4366179611","https://openalex.org/W2029210135"],"abstract_inverted_index":{"Reduced":[0],"precision":[1],"datatypes":[2],"have":[3],"become":[4],"essential":[5],"to":[6,71,124,162],"the":[7,21,25,58,91,110,126,135],"efficient":[8],"training":[9,62,89,127,153],"and":[10,51],"deployment":[11],"of":[12,27,81,112,128],"Deep":[13],"Neural":[14],"Networks":[15],"(DNNs).":[16],"A":[17],"recent":[18],"development":[19],"in":[20],"field":[22],"has":[23,69],"been":[24],"emergence":[26],"block-scaled":[28,82],"datatypes:":[29],"tensor":[30],"representation":[31],"formats":[32,46],"derived":[33],"from":[34],"floating-point,":[35],"that":[36],"share":[37],"a":[38,78],"common":[39],"exponent":[40,114,137],"across":[41],"multiple":[42],"elements.":[43],"While":[44],"these":[45],"are":[47],"being":[48],"broadly":[49],"adopted":[50],"optimised":[52],"for":[53,61,87],"by":[54,132],"DNN-specific":[55],"inference":[56],"accelerators,":[57],"potential":[59],"benefits":[60],"workloads":[63],"on":[64,145],"general-purpose":[65],"(GP)":[66],"vector":[67,96],"processors":[68],"yet":[70],"be":[72],"thoroughly":[73],"explored.":[74],"This":[75,142],"work":[76],"proposes":[77],"benchmarked":[79],"implementation":[80],"general":[83],"matrix":[84],"multiplications":[85],"(GeMM)":[86],"DNN":[88],"at":[90],"edge":[92],"using":[93],"commercially":[94],"available":[95],"instruction":[97],"sets":[98],"(ARM":[99],"SVE).":[100],"Using":[101],"this":[102,122],"implementation,":[103],"we":[104],"highlight":[105],"an":[106],"accuracy-speed":[107],"trade-off":[108],"involving":[109],"shape":[111],"shared":[113,136],"blocks":[115],"-":[116],"vectors":[117],"or":[118],"squares.":[119],"We":[120],"exploit":[121],"result":[123],"optimize":[125],"fully":[129],"connected":[130],"networks":[131],"dynamically":[133],"adapting":[134],"block":[138],"shapes":[139],"during":[140],"training.":[141],"strategy":[143],"yields":[144],"average":[146],"around":[147],"<tex":[148,155],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[149,156],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$1.95":[150],"\\times$</tex>":[151],"faster":[152],"with":[154],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$2\\times$</tex>":[157],"lower":[158],"memory":[159],"footprint":[160],"compared":[161],"standard":[163],"IEEE":[164],"32-bit":[165],"floating":[166],"point":[167],"(FP32),":[168],"while":[169],"achieving":[170],"similar":[171],"accuracy.":[172]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":1}],"updated_date":"2025-12-27T23:08:20.325037","created_date":"2025-10-10T00:00:00"}