{"id":"https://openalex.org/W7117543528","doi":"https://doi.org/10.1109/iccd65941.2025.00040","title":"Dissecting and Re-Architecting 3D NAND Flash PIM Arrays for Efficient Single-Batch Token Generation in LLMS","display_name":"Dissecting and Re-Architecting 3D NAND Flash PIM Arrays for Efficient Single-Batch Token Generation in LLMS","publication_year":2025,"publication_date":"2025-11-10","ids":{"openalex":"https://openalex.org/W7117543528","doi":"https://doi.org/10.1109/iccd65941.2025.00040"},"language":null,"primary_location":{"id":"doi:10.1109/iccd65941.2025.00040","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccd65941.2025.00040","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 43rd International Conference on Computer Design (ICCD)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5036660202","display_name":"Yongjoo Jang","orcid":"https://orcid.org/0000-0001-5009-0559"},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Yongjoo Jang","raw_affiliation_strings":["Korea University,Seoul,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Seoul,South Korea","institution_ids":["https://openalex.org/I197347611"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121558017","display_name":"Sangwoo Hwang","orcid":null},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Sangwoo Hwang","raw_affiliation_strings":["Korea University,Seoul,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Seoul,South Korea","institution_ids":["https://openalex.org/I197347611"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121532311","display_name":"Hojin Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Hojin Lee","raw_affiliation_strings":["Korea University,Seoul,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Seoul,South Korea","institution_ids":["https://openalex.org/I197347611"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121541079","display_name":"Sangwoo Jung","orcid":null},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Sangwoo Jung","raw_affiliation_strings":["Korea University,Seoul,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Seoul,South Korea","institution_ids":["https://openalex.org/I197347611"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121556948","display_name":"Donghun Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Donghun Lee","raw_affiliation_strings":["Korea University,Seoul,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Seoul,South Korea","institution_ids":["https://openalex.org/I197347611"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121554584","display_name":"Wonbo Shim","orcid":null},"institutions":[{"id":"https://openalex.org/I118373667","display_name":"Seoul National University of Science and Technology","ror":"https://ror.org/00chfja07","country_code":"KR","type":"education","lineage":["https://openalex.org/I118373667"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Wonbo Shim","raw_affiliation_strings":["Seoul National University of Science and Technology,Seoul,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Seoul National University of Science and Technology,Seoul,South Korea","institution_ids":["https://openalex.org/I118373667"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5103272968","display_name":"Jaeha Kung","orcid":"https://orcid.org/0000-0002-2027-8531"},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jaeha Kung","raw_affiliation_strings":["Korea University,Seoul,South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Seoul,South Korea","institution_ids":["https://openalex.org/I197347611"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.54383032,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"237","last_page":"244"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.9699000120162964,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.9699000120162964,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.01720000058412552,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10049","display_name":"Magnetic properties of thin films","score":0.0026000000070780516,"subfield":{"id":"https://openalex.org/subfields/3107","display_name":"Atomic and Molecular Physics, and Optics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/nand-gate","display_name":"NAND gate","score":0.7281000018119812},{"id":"https://openalex.org/keywords/dram","display_name":"Dram","score":0.6920999884605408},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.5954999923706055},{"id":"https://openalex.org/keywords/flash","display_name":"Flash (photography)","score":0.48809999227523804},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.4715000092983246},{"id":"https://openalex.org/keywords/cas-latency","display_name":"CAS latency","score":0.4668000042438507}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7526999711990356},{"id":"https://openalex.org/C124296912","wikidata":"https://www.wikidata.org/wiki/Q575178","display_name":"NAND gate","level":3,"score":0.7281000018119812},{"id":"https://openalex.org/C7366592","wikidata":"https://www.wikidata.org/wiki/Q1255620","display_name":"Dram","level":2,"score":0.6920999884605408},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5954999923706055},{"id":"https://openalex.org/C2777526259","wikidata":"https://www.wikidata.org/wiki/Q221836","display_name":"Flash (photography)","level":2,"score":0.48809999227523804},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.4715000092983246},{"id":"https://openalex.org/C189930140","wikidata":"https://www.wikidata.org/wiki/Q1112878","display_name":"CAS latency","level":4,"score":0.4668000042438507},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.46639999747276306},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.46309998631477356},{"id":"https://openalex.org/C136085584","wikidata":"https://www.wikidata.org/wiki/Q910289","display_name":"Overlay","level":2,"score":0.4510999917984009},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.4489000141620636},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.3562000095844269},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3312999904155731},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.32330000400543213},{"id":"https://openalex.org/C2776531357","wikidata":"https://www.wikidata.org/wiki/Q174077","display_name":"Flash memory","level":2,"score":0.31369999051094055},{"id":"https://openalex.org/C2779602883","wikidata":"https://www.wikidata.org/wiki/Q15544750","display_name":"Memory architecture","level":2,"score":0.2660999894142151},{"id":"https://openalex.org/C195053848","wikidata":"https://www.wikidata.org/wiki/Q7894141","display_name":"Universal memory","level":5,"score":0.2646999955177307}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iccd65941.2025.00040","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccd65941.2025.00040","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 43rd International Conference on Computer Design (ICCD)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1612713835","display_name":null,"funder_award_id":"RS-2023-00258227","funder_id":"https://openalex.org/F4320322120","funder_display_name":"National Research Foundation of Korea"},{"id":"https://openalex.org/G2463336223","display_name":null,"funder_award_id":"RS-2023-00229849;","funder_id":"https://openalex.org/F4320328359","funder_display_name":"Ministry of Science and ICT, South Korea"},{"id":"https://openalex.org/G3404986031","display_name":null,"funder_award_id":"2022-0-01170","funder_id":"https://openalex.org/F4320335839","funder_display_name":"National Institute of Information and Communications Technology"}],"funders":[{"id":"https://openalex.org/F4320322120","display_name":"National Research Foundation of Korea","ror":"https://ror.org/013aysd81"},{"id":"https://openalex.org/F4320328359","display_name":"Ministry of Science and ICT, South Korea","ror":"https://ror.org/01wpjm123"},{"id":"https://openalex.org/F4320335839","display_name":"National Institute of Information and Communications Technology","ror":"https://ror.org/016bgq349"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":13,"referenced_works":["https://openalex.org/W1608061514","https://openalex.org/W2782046614","https://openalex.org/W2790558670","https://openalex.org/W2899332646","https://openalex.org/W2951313701","https://openalex.org/W3036619909","https://openalex.org/W3114125346","https://openalex.org/W3190456202","https://openalex.org/W4308083744","https://openalex.org/W4308083842","https://openalex.org/W4360831773","https://openalex.org/W4387321091","https://openalex.org/W4394998968"],"related_works":[],"abstract_inverted_index":{"The":[0],"advancement":[1],"of":[2,12],"large":[3],"language":[4],"models":[5,9,22],"has":[6],"led":[7],"to":[8,29,48,61,117],"with":[10,79,90,112,120],"billions":[11],"parameters,":[13],"significantly":[14],"increasing":[15],"memory":[16,147],"and":[17,33,73,86,100,114],"compute":[18],"demands.":[19],"Serving":[20],"such":[21],"on":[23],"conventional":[24],"hardware":[25],"is":[26],"challenging":[27],"due":[28],"limited":[30],"DRAM":[31,64],"capacity":[32,65],"high":[34,58],"GPU":[35],"costs.":[36],"Thus,":[37],"in":[38],"this":[39],"work,":[40],"we":[41,96],"propose":[42],"offloading":[43],"the":[44,63,91,131,146],"single-batch":[45],"token":[46],"generation":[47],"a":[49,75],"3D":[50,69,133],"NAND":[51,70,134],"flash":[52,71,135],"processing-in-memory":[53],"(PIM)":[54],"device,":[55],"leveraging":[56],"its":[57],"storage":[59],"density":[60],"overcome":[62],"wall.":[66],"We":[67],"explore":[68],"configurations":[72],"present":[74],"re-architected":[76],"PIM":[77,93,136],"array":[78,94],"an":[80],"H-tree":[81],"network":[82],"for":[83,103],"optimal":[84],"latency":[85,123],"cell":[87],"density.":[88],"Along":[89],"well-chosen":[92],"size,":[95],"develop":[97],"operation":[98],"tiling":[99],"mapping":[101],"methods":[102],"LLM":[104],"layers,":[105],"achieving":[106],"a$2.4":[107],"\\times$speedup":[108],"over":[109],"four":[110,118],"RTX4090":[111],"vLLM":[113],"comparable":[115],"performance":[116],"A100":[119],"only":[121],"4.9%":[122],"overhead.":[124,152],"Our":[125],"detailed":[126],"area":[127,144,151],"analysis":[128],"reveals":[129],"that":[130],"proposed":[132],"architecture":[137],"can":[138],"be":[139],"integrated":[140],"within":[141],"a$4.98":[142],"~\\text{mm}^{2}$die":[143],"under":[145],"array,":[148],"without":[149],"extra":[150]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-12-30T00:00:00"}
