{"id":"https://openalex.org/W7133490529","doi":"https://doi.org/10.1109/hpca68181.2026.11408592","title":"PIMphony: Overcoming Bandwidth and Capacity Inefficiency in PIM-Based Long-Context LLM Inference System","display_name":"PIMphony: Overcoming Bandwidth and Capacity Inefficiency in PIM-Based Long-Context LLM Inference System","publication_year":2026,"publication_date":"2026-01-31","ids":{"openalex":"https://openalex.org/W7133490529","doi":"https://doi.org/10.1109/hpca68181.2026.11408592"},"language":null,"primary_location":{"id":"doi:10.1109/hpca68181.2026.11408592","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpca68181.2026.11408592","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)","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/A5044511282","display_name":"Hyucksung Kwon","orcid":null},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":true,"raw_author_name":"Hyucksung Kwon","raw_affiliation_strings":["Hanyang University,Seoul,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I4575257"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128065573","display_name":"Kyungmo Koo","orcid":null},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Kyungmo Koo","raw_affiliation_strings":["Hanyang University,Seoul,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I4575257"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041027404","display_name":"Junsu Kim","orcid":"https://orcid.org/0000-0001-9583-7978"},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]},{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Janghyeon Kim","raw_affiliation_strings":["Hanyang University,Seoul,Republic of Korea","KAIST,Daejeon,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I4575257"]},{"raw_affiliation_string":"KAIST,Daejeon,Republic of Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128104355","display_name":"Woongkyu Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Woongkyu Lee","raw_affiliation_strings":["Hanyang University,Seoul,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I4575257"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128075408","display_name":"Minjae Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Minjae Lee","raw_affiliation_strings":["Hanyang University,Seoul,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I4575257"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Gyeonggeun Jung","orcid":null},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Gyeonggeun Jung","raw_affiliation_strings":["KAIST,Daejeon,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"KAIST,Daejeon,Republic of Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Hyungdeok Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hyungdeok Lee","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128090713","display_name":"Yousub Jung","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yousub Jung","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128068993","display_name":"Jaehan Park","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jaehan Park","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113274482","display_name":"Yosub Song","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yosub Song","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128057396","display_name":"Byeongsu Yang","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Byeongsu Yang","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Haerang Choi","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Haerang Choi","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128063813","display_name":"Guhyun Kim","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Guhyun Kim","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128078595","display_name":"Jongsoon Won","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jongsoon Won","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027083700","display_name":"Woojae Shin","orcid":"https://orcid.org/0000-0002-6155-3712"},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Woojae Shin","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100600387","display_name":"Changhyun Kim","orcid":"https://orcid.org/0000-0001-7181-136X"},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Changhyun Kim","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128112528","display_name":"Gyeongcheol Shin","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Gyeongcheol Shin","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yongkee Kwon","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yongkee Kwon","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112662441","display_name":"Ilkon Kim","orcid":null},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ilkon Kim","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039338759","display_name":"Euicheol Lim","orcid":"https://orcid.org/0000-0002-8910-533X"},"institutions":[{"id":"https://openalex.org/I10654025","display_name":"SK Group (United States)","ror":"https://ror.org/00qajw440","country_code":"US","type":"company","lineage":["https://openalex.org/I10654025","https://openalex.org/I134353371"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Euicheol Lim","raw_affiliation_strings":["Solution Advanced Technology, SK hynix,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Solution Advanced Technology, SK hynix,Republic of Korea","institution_ids":["https://openalex.org/I10654025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010309237","display_name":"J. Dongun Kim","orcid":"https://orcid.org/0000-0003-0462-6521"},"institutions":[{"id":"https://openalex.org/I157485424","display_name":"Korea Advanced Institute of Science and Technology","ror":"https://ror.org/05apxxy63","country_code":"KR","type":"education","lineage":["https://openalex.org/I157485424"]},{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"John Kim","raw_affiliation_strings":["Hanyang University,Seoul,Republic of Korea","KAIST,Daejeon,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I4575257"]},{"raw_affiliation_string":"KAIST,Daejeon,Republic of Korea","institution_ids":["https://openalex.org/I157485424"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101935712","display_name":"Jungwook Choi","orcid":"https://orcid.org/0000-0002-5691-4771"},"institutions":[{"id":"https://openalex.org/I4575257","display_name":"Hanyang University","ror":"https://ror.org/046865y68","country_code":"KR","type":"education","lineage":["https://openalex.org/I4575257"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jungwook Choi","raw_affiliation_strings":["Hanyang University,Seoul,Republic of Korea"],"affiliations":[{"raw_affiliation_string":"Hanyang University,Seoul,Republic of Korea","institution_ids":["https://openalex.org/I4575257"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":22,"corresponding_author_ids":["https://openalex.org/A5044511282"],"corresponding_institution_ids":["https://openalex.org/I4575257"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.50041326,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"21"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11873","display_name":"PAPR reduction in OFDM","score":0.09960000216960907,"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"}},"topics":[{"id":"https://openalex.org/T11873","display_name":"PAPR reduction in OFDM","score":0.09960000216960907,"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/T12131","display_name":"Wireless Signal Modulation Classification","score":0.09149999916553497,"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"}},{"id":"https://openalex.org/T12791","display_name":"Full-Duplex Wireless Communications","score":0.05700000002980232,"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/inefficiency","display_name":"Inefficiency","score":0.5041000247001648},{"id":"https://openalex.org/keywords/bandwidth","display_name":"Bandwidth (computing)","score":0.4693000018596649},{"id":"https://openalex.org/keywords/capacity-planning","display_name":"Capacity planning","score":0.4083999991416931},{"id":"https://openalex.org/keywords/control-system","display_name":"Control system","score":0.3409000039100647},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.3370000123977661},{"id":"https://openalex.org/keywords/control-theory","display_name":"Control theory (sociology)","score":0.33410000801086426}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5849000215530396},{"id":"https://openalex.org/C2778869765","wikidata":"https://www.wikidata.org/wiki/Q6028363","display_name":"Inefficiency","level":2,"score":0.5041000247001648},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.4693000018596649},{"id":"https://openalex.org/C2781007418","wikidata":"https://www.wikidata.org/wiki/Q1456934","display_name":"Capacity planning","level":2,"score":0.4083999991416931},{"id":"https://openalex.org/C17500928","wikidata":"https://www.wikidata.org/wiki/Q959968","display_name":"Control system","level":2,"score":0.3409000039100647},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3370000123977661},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.33410000801086426},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.2808000147342682},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2687000036239624},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.26759999990463257},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.26739999651908875},{"id":"https://openalex.org/C119247159","wikidata":"https://www.wikidata.org/wiki/Q1366192","display_name":"System identification","level":3,"score":0.26739999651908875},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.2632000148296356}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpca68181.2026.11408592","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpca68181.2026.11408592","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":35,"referenced_works":["https://openalex.org/W2725159389","https://openalex.org/W3100710793","https://openalex.org/W3132046658","https://openalex.org/W3134274954","https://openalex.org/W3189166979","https://openalex.org/W4221001402","https://openalex.org/W4226126604","https://openalex.org/W4280496502","https://openalex.org/W4281733247","https://openalex.org/W4285121610","https://openalex.org/W4295036115","https://openalex.org/W4297097348","https://openalex.org/W4308083513","https://openalex.org/W4312697304","https://openalex.org/W4313289674","https://openalex.org/W4387064011","https://openalex.org/W4387064057","https://openalex.org/W4387321091","https://openalex.org/W4389476267","https://openalex.org/W4392427708","https://openalex.org/W4393145094","https://openalex.org/W4394998968","https://openalex.org/W4394999021","https://openalex.org/W4395073431","https://openalex.org/W4395106409","https://openalex.org/W4399402229","https://openalex.org/W4401211806","https://openalex.org/W4402475900","https://openalex.org/W4407425079","https://openalex.org/W4408323861","https://openalex.org/W4408925909","https://openalex.org/W4409248689","https://openalex.org/W4409328517","https://openalex.org/W4410537450","https://openalex.org/W4411486169"],"related_works":[],"abstract_inverted_index":{"The":[0],"expansion":[1],"of":[2,76,143,159],"long-context":[3,127,163],"Large":[4],"Language":[5],"Models":[6],"(LLMs)":[7],"creates":[8],"significant":[9],"memory":[10,41,104,109],"system":[11],"challenges.":[12],"While":[13],"Processing-in-Memory":[14],"(PIM)":[15],"is":[16],"a":[17,54,96,119],"promising":[18],"accelerator,":[19],"we":[20,51],"identify":[21],"that":[22,57],"it":[23],"suffers":[24],"from":[25,43],"critical":[26],"inefficiencies":[27],"when":[28],"scaled":[29],"to":[30,106,131,145],"long":[31],"contexts:":[32],"severe":[33],"channel":[34,73],"underutilization,":[35],"performancelimiting":[36],"I/O":[37,87],"bottlenecks,":[38],"and":[39,93,116,134,150],"massive":[40],"waste":[42],"static":[44,108],"KV":[45],"cache":[46],"management.":[47],"In":[48],"this":[49],"work,":[50],"propose":[52],"PIMphony,":[53],"PIM":[55,68,81,98],"orchestrator":[56],"systematically":[58],"resolves":[59],"these":[60],"issues":[61],"with":[62],"three":[63],"co-designed":[64],"techniques.":[65],"First,":[66],"Token-Centric":[67],"Partitioning":[69],"(TCP)":[70],"ensures":[71],"high":[72],"utilization":[74],"regardless":[75],"batch":[77],"size.":[78],"Second,":[79],"Dynamic":[80,97],"Command":[82],"Scheduling":[83],"(DCS)":[84],"mitigates":[85],"the":[86],"bottleneck":[88],"by":[89],"overlapping":[90],"data":[91],"movement":[92],"computation.":[94],"Finally,":[95],"Access":[99],"(DPA)":[100],"controller":[101],"enables":[102],"dynamic":[103],"management":[105],"eliminate":[107],"waste.":[110],"Implemented":[111],"via":[112],"an":[113],"MLIR-based":[114],"compiler":[115],"evaluated":[117],"on":[118,147,152],"cycle-accurate":[120],"simulator,":[121],"PIMphony":[122],"significantly":[123],"improves":[124],"throughput":[125],"for":[126],"LLM":[128],"inference":[129],"(up":[130],"72B":[132],"parameters":[133],"1M":[135],"context":[136],"length).":[137],"Our":[138],"evaluations":[139],"show":[140],"performance":[141],"boosts":[142],"up":[144],"11.3\u00d7":[146],"PIM-only":[148],"systems":[149],"8.4\u00d7":[151],"xPU+PIM":[153],"systems,":[154],"enabling":[155],"more":[156],"efficient":[157],"deployment":[158],"LLMs":[160],"in":[161],"real-world":[162],"applications.":[164]},"counts_by_year":[],"updated_date":"2026-04-16T08:26:57.006410","created_date":"2026-03-05T00:00:00"}
