{"id":"https://openalex.org/W4399282321","doi":"https://doi.org/10.1145/3650200.3656628","title":"An Autonomous Parallelization of Transformer Model Inference on Heterogeneous Edge Devices","display_name":"An Autonomous Parallelization of Transformer Model Inference on Heterogeneous Edge Devices","publication_year":2024,"publication_date":"2024-05-30","ids":{"openalex":"https://openalex.org/W4399282321","doi":"https://doi.org/10.1145/3650200.3656628"},"language":"en","primary_location":{"id":"doi:10.1145/3650200.3656628","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3650200.3656628","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3650200.3656628","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 38th ACM International Conference on Supercomputing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3650200.3656628","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5074595296","display_name":"Juhyeon Lee","orcid":"https://orcid.org/0009-0008-2042-1159"},"institutions":[{"id":"https://openalex.org/I148751991","display_name":"Sogang University","ror":"https://ror.org/056tn4839","country_code":"KR","type":"education","lineage":["https://openalex.org/I148751991"]}],"countries":["KR"],"is_corresponding":true,"raw_author_name":"Juhyeon Lee","raw_affiliation_strings":["Sogang University, Korea, South (Republic of Korea)"],"raw_orcid":"https://orcid.org/0009-0008-2042-1159","affiliations":[{"raw_affiliation_string":"Sogang University, Korea, South (Republic of Korea)","institution_ids":["https://openalex.org/I148751991"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5098993311","display_name":"Insung Bahk","orcid":"https://orcid.org/0009-0001-0177-4636"},"institutions":[{"id":"https://openalex.org/I146824383","display_name":"Hallym University","ror":"https://ror.org/03sbhge02","country_code":"KR","type":"education","lineage":["https://openalex.org/I146824383"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Insung Bahk","raw_affiliation_strings":["Hallym University, Korea, South (Republic of Korea)"],"raw_orcid":"https://orcid.org/0009-0001-0177-4636","affiliations":[{"raw_affiliation_string":"Hallym University, Korea, South (Republic of Korea)","institution_ids":["https://openalex.org/I146824383"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065198808","display_name":"Hoseung Kim","orcid":"https://orcid.org/0009-0005-1454-3617"},"institutions":[{"id":"https://openalex.org/I848706","display_name":"Sungkyunkwan University","ror":"https://ror.org/04q78tk20","country_code":"KR","type":"education","lineage":["https://openalex.org/I848706"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Hoseung Kim","raw_affiliation_strings":["Sungkyunkwan University, Republic of Korea"],"raw_orcid":"https://orcid.org/0009-0005-1454-3617","affiliations":[{"raw_affiliation_string":"Sungkyunkwan University, Republic of Korea","institution_ids":["https://openalex.org/I848706"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073916008","display_name":"S.-J. Jeong","orcid":"https://orcid.org/0000-0002-6016-1696"},"institutions":[{"id":"https://openalex.org/I4921948","display_name":"Pusan National University","ror":"https://ror.org/01an57a31","country_code":"KR","type":"education","lineage":["https://openalex.org/I4921948"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Sinjin Jeong","raw_affiliation_strings":["Pusan University, Republic of Korea"],"raw_orcid":"https://orcid.org/0000-0002-6016-1696","affiliations":[{"raw_affiliation_string":"Pusan University, Republic of Korea","institution_ids":["https://openalex.org/I4921948"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5089446956","display_name":"S. Lee","orcid":"https://orcid.org/0000-0002-5526-6127"},"institutions":[{"id":"https://openalex.org/I130701444","display_name":"Georgia Institute of Technology","ror":"https://ror.org/01zkghx44","country_code":"US","type":"education","lineage":["https://openalex.org/I130701444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Suyeon Lee","raw_affiliation_strings":["Georgia Institute of Technology, United States of America"],"raw_orcid":"https://orcid.org/0000-0002-5526-6127","affiliations":[{"raw_affiliation_string":"Georgia Institute of Technology, United States of America","institution_ids":["https://openalex.org/I130701444"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5005461365","display_name":"Donghyun Min","orcid":"https://orcid.org/0000-0002-6043-9264"},"institutions":[{"id":"https://openalex.org/I148751991","display_name":"Sogang University","ror":"https://ror.org/056tn4839","country_code":"KR","type":"education","lineage":["https://openalex.org/I148751991"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Donghyun Min","raw_affiliation_strings":["Sogang University, Korea, South (Republic of Korea)"],"raw_orcid":"https://orcid.org/0000-0002-6043-9264","affiliations":[{"raw_affiliation_string":"Sogang University, Korea, South (Republic of Korea)","institution_ids":["https://openalex.org/I148751991"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5074595296"],"corresponding_institution_ids":["https://openalex.org/I148751991"],"apc_list":null,"apc_paid":null,"fwci":1.4202,"has_fulltext":false,"cited_by_count":7,"citation_normalized_percentile":{"value":0.78417559,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"50","last_page":"61"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12169","display_name":"Non-Destructive Testing Techniques","score":0.996399998664856,"subfield":{"id":"https://openalex.org/subfields/2210","display_name":"Mechanical 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/T12169","display_name":"Non-Destructive Testing Techniques","score":0.996399998664856,"subfield":{"id":"https://openalex.org/subfields/2210","display_name":"Mechanical 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/T11343","display_name":"Power Transformer Diagnostics and Insulation","score":0.9922000169754028,"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/T10320","display_name":"Neural Networks and Applications","score":0.9868000149726868,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7512619495391846},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6544270515441895},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.5939158797264099},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.5676084756851196},{"id":"https://openalex.org/keywords/automatic-parallelization","display_name":"Automatic parallelization","score":0.4722824990749359},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.4375390410423279},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.2391919493675232},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.15119868516921997},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.09008225798606873},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.06547823548316956}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7512619495391846},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6544270515441895},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5939158797264099},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5676084756851196},{"id":"https://openalex.org/C164833996","wikidata":"https://www.wikidata.org/wiki/Q2323839","display_name":"Automatic parallelization","level":3,"score":0.4722824990749359},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.4375390410423279},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.2391919493675232},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.15119868516921997},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.09008225798606873},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.06547823548316956},{"id":"https://openalex.org/C169590947","wikidata":"https://www.wikidata.org/wiki/Q47506","display_name":"Compiler","level":2,"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:10.1145/3650200.3656628","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3650200.3656628","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3650200.3656628","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 38th ACM International Conference on Supercomputing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3650200.3656628","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3650200.3656628","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3650200.3656628","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 38th ACM International Conference on Supercomputing","raw_type":"proceedings-article"},"sustainable_development_goals":[{"display_name":"Decent work and economic growth","score":0.5899999737739563,"id":"https://metadata.un.org/sdg/8"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":false},"content_urls":{"pdf":"https://content.openalex.org/works/W4399282321.pdf"},"referenced_works_count":23,"referenced_works":["https://openalex.org/W2091432990","https://openalex.org/W2172620437","https://openalex.org/W2604319603","https://openalex.org/W2612193523","https://openalex.org/W2896180420","https://openalex.org/W2950865323","https://openalex.org/W2960833983","https://openalex.org/W2970971581","https://openalex.org/W2980856918","https://openalex.org/W2983440318","https://openalex.org/W3005276590","https://openalex.org/W3023935494","https://openalex.org/W3081168214","https://openalex.org/W3110777925","https://openalex.org/W3111535274","https://openalex.org/W3138516171","https://openalex.org/W3153465022","https://openalex.org/W3177372937","https://openalex.org/W3204998121","https://openalex.org/W4221146321","https://openalex.org/W4224314467","https://openalex.org/W4283204791","https://openalex.org/W4313496563"],"related_works":["https://openalex.org/W2555203579","https://openalex.org/W1842571235","https://openalex.org/W2994999011","https://openalex.org/W3024114208","https://openalex.org/W1985416522","https://openalex.org/W2911340114","https://openalex.org/W4206178588","https://openalex.org/W3094491777","https://openalex.org/W3214715529","https://openalex.org/W4233032676"],"abstract_inverted_index":{"The":[0],"utilization":[1],"of":[2,16,54,77,94,109,143,156],"advancing":[3],"transformer-based":[4,70,95],"deep":[5],"neural":[6],"network":[7,127,147],"(DNN)":[8],"models":[9,25,71,97],"in":[10,69],"edge":[11,27,78,100,144],"environments":[12],"holds":[13],"the":[14,74,136,140,162,171],"promise":[15],"improving":[17],"productivity":[18],"for":[19],"intelligent":[20],"tasks.":[21],"However,":[22],"deploying":[23],"these":[24],"on":[26,98,170],"devices":[28,45,145],"with":[29],"limited":[30],"resources":[31],"encounters":[32],"significant":[33],"performance":[34,154],"challenges.":[35],"Previous":[36],"solutions":[37],"have":[38],"attempted":[39],"to":[40,65,90,113,123,125,161],"distribute":[41],"computation":[42],"tasks":[43],"across":[44],"and":[46,72,118,128,146,158,166],"perform":[47],"parallel":[48,92,116],"inferences":[49],"but":[50],"often":[51],"fall":[52],"short":[53],"meeting":[55],"service-level":[56],"objectives":[57],"(SLO).":[58],"This":[59],"limitation":[60],"arises":[61],"from":[62],"their":[63],"inability":[64],"effectively":[66],"harness":[67],"parallelization":[68],"consider":[73],"resource":[75,130,141],"diversity":[76,142],"devices.":[79],"In":[80],"this":[81],"paper,":[82],"we":[83],"propose":[84],"Hepti,":[85],"a":[86,152],"practical":[87],"framework":[88],"designed":[89],"facilitate":[91],"inference":[93,117,164],"DNN":[96],"heterogeneous":[99],"environments.":[101],"Hepti":[102,137,150],"is":[103],"armed":[104],"with:":[105],"1)":[106],"an":[107],"understanding":[108],"transformer":[110],"model":[111,168],"architecture":[112],"enable":[114],"effective":[115],"2)":[119],"dynamic":[120],"workload":[121],"optimization":[122],"adapt":[124],"changing":[126],"device":[129],"capabilities.":[131],"Our":[132],"evaluations":[133],"confirmed":[134],"that":[135],"autonomously":[138],"assesses":[139],"status.":[148],"Furthermore,":[149],"achieves":[151],"maximum":[153],"improvement":[155],"49.1%":[157],"37.1%":[159],"compared":[160],"local":[163],"approach":[165],"state-of-the-art":[167],"parallelisms":[169],"BERT-Large":[172],"model.":[173]},"counts_by_year":[{"year":2025,"cited_by_count":7}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}