{"id":"https://openalex.org/W4413755433","doi":"https://doi.org/10.1109/isvlsi65124.2025.11130222","title":"CIM for Transformer Models: Enhancing Large Language Model Inference Efficiency","display_name":"CIM for Transformer Models: Enhancing Large Language Model Inference Efficiency","publication_year":2025,"publication_date":"2025-07-06","ids":{"openalex":"https://openalex.org/W4413755433","doi":"https://doi.org/10.1109/isvlsi65124.2025.11130222"},"language":"en","primary_location":{"id":"doi:10.1109/isvlsi65124.2025.11130222","is_oa":false,"landing_page_url":"https://doi.org/10.1109/isvlsi65124.2025.11130222","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/A5100339707","display_name":"Miao Li","orcid":"https://orcid.org/0000-0002-5501-2856"},"institutions":[{"id":"https://openalex.org/I22265921","display_name":"National Central University","ror":"https://ror.org/00944ve71","country_code":"TW","type":"education","lineage":["https://openalex.org/I22265921"]}],"countries":["TW"],"is_corresponding":true,"raw_author_name":"Meng-Syuan Li","raw_affiliation_strings":["National Central University,Department of Electrical Engineering,Taoyuan,Taiwan"],"affiliations":[{"raw_affiliation_string":"National Central University,Department of Electrical Engineering,Taoyuan,Taiwan","institution_ids":["https://openalex.org/I22265921"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109657626","display_name":"Junxin Ke","orcid":null},"institutions":[{"id":"https://openalex.org/I22265921","display_name":"National Central University","ror":"https://ror.org/00944ve71","country_code":"TW","type":"education","lineage":["https://openalex.org/I22265921"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Jung-Fang Ke","raw_affiliation_strings":["National Central University,Department of Electrical Engineering,Taoyuan,Taiwan"],"affiliations":[{"raw_affiliation_string":"National Central University,Department of Electrical Engineering,Taoyuan,Taiwan","institution_ids":["https://openalex.org/I22265921"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067971879","display_name":"En-Ming Huang","orcid":"https://orcid.org/0000-0003-2196-2834"},"institutions":[{"id":"https://openalex.org/I16733864","display_name":"National Taiwan University","ror":"https://ror.org/05bqach95","country_code":"TW","type":"education","lineage":["https://openalex.org/I16733864"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"En-Ming Huang","raw_affiliation_strings":["National Taiwan University,Department of Computer Science,Taipei,Taiwan"],"affiliations":[{"raw_affiliation_string":"National Taiwan University,Department of Computer Science,Taipei,Taiwan","institution_ids":["https://openalex.org/I16733864"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100321247","display_name":"Zhiwei Liu","orcid":"https://orcid.org/0000-0003-1525-1067"},"institutions":[{"id":"https://openalex.org/I22265921","display_name":"National Central University","ror":"https://ror.org/00944ve71","country_code":"TW","type":"education","lineage":["https://openalex.org/I22265921"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Zhi-Wei Liu","raw_affiliation_strings":["National Central University,Department of Electrical Engineering,Taoyuan,Taiwan"],"affiliations":[{"raw_affiliation_string":"National Central University,Department of Electrical Engineering,Taoyuan,Taiwan","institution_ids":["https://openalex.org/I22265921"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5011277985","display_name":"Yuguang Chen","orcid":"https://orcid.org/0000-0003-4520-5395"},"institutions":[{"id":"https://openalex.org/I22265921","display_name":"National Central University","ror":"https://ror.org/00944ve71","country_code":"TW","type":"education","lineage":["https://openalex.org/I22265921"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Yu-Guang Chen","raw_affiliation_strings":["National Central University,Department of Electrical Engineering,Taoyuan,Taiwan"],"affiliations":[{"raw_affiliation_string":"National Central University,Department of Electrical Engineering,Taoyuan,Taiwan","institution_ids":["https://openalex.org/I22265921"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101593316","display_name":"Chun\u2010Yi Lee","orcid":"https://orcid.org/0000-0002-5481-9260"},"institutions":[{"id":"https://openalex.org/I16733864","display_name":"National Taiwan University","ror":"https://ror.org/05bqach95","country_code":"TW","type":"education","lineage":["https://openalex.org/I16733864"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Chun-Yi Lee","raw_affiliation_strings":["National Taiwan University,Department of Computer Science,Taipei,Taiwan"],"affiliations":[{"raw_affiliation_string":"National Taiwan University,Department of Computer Science,Taipei,Taiwan","institution_ids":["https://openalex.org/I16733864"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5100339707"],"corresponding_institution_ids":["https://openalex.org/I22265921"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.12995044,"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":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9578999876976013,"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"}},"topics":[{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9578999876976013,"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/T10028","display_name":"Topic Modeling","score":0.904699981212616,"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/inference","display_name":"Inference","score":0.7154659032821655},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.6812266111373901},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6722531318664551},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.28637129068374634},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.1408553421497345},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.1008962094783783},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.07756230235099792}],"concepts":[{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7154659032821655},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6812266111373901},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6722531318664551},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.28637129068374634},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.1408553421497345},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.1008962094783783},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.07756230235099792}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/isvlsi65124.2025.11130222","is_oa":false,"landing_page_url":"https://doi.org/10.1109/isvlsi65124.2025.11130222","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"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320331164","display_name":"National Science and Technology Council","ror":"https://ror.org/00wnb9798"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":19,"referenced_works":["https://openalex.org/W2010202670","https://openalex.org/W2117422822","https://openalex.org/W2168249605","https://openalex.org/W2508602506","https://openalex.org/W2518281301","https://openalex.org/W2536878608","https://openalex.org/W2613989746","https://openalex.org/W2768489488","https://openalex.org/W2785141883","https://openalex.org/W2818007000","https://openalex.org/W2985844759","https://openalex.org/W3018945530","https://openalex.org/W3111375540","https://openalex.org/W3158997543","https://openalex.org/W3201198776","https://openalex.org/W4220702691","https://openalex.org/W4286571874","https://openalex.org/W4321636575","https://openalex.org/W4385245566"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4206178588","https://openalex.org/W3094491777","https://openalex.org/W3214715529","https://openalex.org/W4287635093"],"abstract_inverted_index":{"In":[0,53],"the":[1,9,41,45,54,68,72,81,87,102,114,122,138,173],"field":[2],"of":[3,56,117,140,175],"large":[4],"language":[5],"model":[6],"(LLM)":[7],"inference,":[8,146],"high":[10],"computational":[11],"demand":[12],"and":[13,19,76,104,112,120,147,168,178,194],"extensive":[14],"memory":[15],"requirements":[16],"for":[17,74,89,171],"weights":[18],"key-value":[20],"(KV)":[21],"cache":[22,70],"storage":[23,75],"present":[24,163],"significant":[25],"challenges.":[26],"This":[27,134],"issue":[28],"becomes":[29],"especially":[30],"problematic":[31],"when":[32],"relying":[33],"exclusively":[34],"on":[35,101,125],"GPUs,":[36,63],"as":[37,130,200],"they":[38],"often":[39],"lack":[40],"capacity":[42],"to":[43,71,86,157,202],"accommodate":[44],"entire":[46],"KV":[47,69,110],"cache,":[48],"particularly":[49],"in":[50,144,187],"larger":[51],"LLMs.":[52],"absence":[55],"direct":[57,115],"communications":[58],"like":[59],"NVlink":[60],"among":[61],"multiple":[62],"LLMs":[64],"typically":[65],"require":[66],"offloading":[67],"CPU":[73,103],"computation,":[77],"followed":[78],"by":[79],"transferring":[80],"multi-head":[82],"attention":[83,95,118,176],"results":[84,184],"back":[85],"GPU":[88],"subsequent":[90],"transformer":[91],"computations.":[92],"Given":[93],"that":[94,152],"score":[96],"computation":[97,116,174],"is":[98,136],"computationally":[99],"demanding":[100],"requires":[105],"substantial":[106],"data":[107],"movement":[108],"between":[109],"caches":[111],"memory,":[113],"scores":[119,177],"even":[121],"feedforward":[123],"layers":[124,180],"Compute-in-Memory":[126],"(CIM)":[127],"systems":[128],"emerges":[129],"a":[131,164,203],"viable":[132],"alternative.":[133],"paper":[135],"at":[137],"forefront":[139],"integrating":[141],"CIM":[142],"technology":[143,156],"LLM":[145],"proposes":[148],"an":[149],"innovative":[150],"architecture":[151],"leverages":[153],"this":[154],"emerging":[155],"enhance":[158],"inference":[159,192],"efficiency.":[160],"Specifically,":[161],"we":[162],"tailored":[165],"CIM-based":[166],"dataflow":[167],"hierarchy":[169],"design":[170],"optimize":[172],"feed-forward":[179],"using":[181],"CIMs.":[182],"The":[183],"show":[185],"improvements":[186],"performance,":[188],"with":[189],"$0.026":[190],"\\times$":[191,198],"latency":[193],"$1.199":[195],"\\times":[196],"10^{-3}":[197],"energy":[199],"compared":[201],"CPU-based":[204],"implementation.":[205]},"counts_by_year":[],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}