{"id":"https://openalex.org/W7204952095","doi":"https://doi.org/10.1109/lca.2026.3729746","title":"FLINT: Efficiently Exploiting High-Bandwidth Flash for Capacity-Scalable LLM Inference Acceleration","display_name":"FLINT: Efficiently Exploiting High-Bandwidth Flash for Capacity-Scalable LLM Inference Acceleration","publication_year":2026,"publication_date":"2026-07-01","ids":{"openalex":"https://openalex.org/W7204952095","doi":"https://doi.org/10.1109/lca.2026.3729746"},"language":"en","primary_location":{"id":"doi:10.1109/lca.2026.3729746","doi":"https://doi.org/10.1109/lca.2026.3729746","is_oa":true,"landing_page_url":"https://doi.org/10.1109/lca.2026.3729746","pdf_url":null,"source":{"id":"https://openalex.org/S17643076","display_name":"IEEE Computer Architecture Letters","issn_l":"1556-6056","issn":["1556-6056","1556-6064","2473-2575"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core","dongbi-b","norway-1","poland-70","russia-white-list-2"],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Computer Architecture Letters","raw_type":"journal-article"},"volume":null,"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.1109/lca.2026.3729746","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5091016237","display_name":"Geraldo Francisco de Oliveira Junior","orcid":"https://orcid.org/0000-0003-1557-4819","observed_orcids":["https://orcid.org/0000-0003-1557-4819"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Geraldo F. Oliveira","raw_affiliation_strings":["Computing Systems Lab, Huawei Zurich Research Center"],"raw_orcid":"https://orcid.org/0000-0003-1557-4819","affiliations":[{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148468537","display_name":"Xiangyu Zhu","orcid":"https://orcid.org/0009-0009-3901-2088","observed_orcids":["https://orcid.org/0009-0009-3901-2088"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiangyu Zhu","raw_affiliation_strings":["Computing Systems Lab, Huawei Zurich Research Center"],"raw_orcid":"https://orcid.org/0009-0009-3901-2088","affiliations":[{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047647989","display_name":"Ahmet Caner Y\u00fcz\u00fcg\u00fcler","orcid":"https://orcid.org/0000-0001-7809-9897","observed_orcids":["https://orcid.org/0000-0001-7809-9897"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ahmet Caner Y\u00fcz\u00fcg\u00fcler","raw_affiliation_strings":["Computing Systems Lab, Huawei Zurich Research Center"],"raw_orcid":"https://orcid.org/0000-0001-7809-9897","affiliations":[{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5116819642","display_name":"Vamanan Arulchelvan","orcid":"https://orcid.org/0009-0008-5623-2376","observed_orcids":["https://orcid.org/0009-0008-5623-2376"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Vamanan Arulchelvan","raw_affiliation_strings":["Computing Systems Lab, Huawei Zurich Research Center"],"raw_orcid":"https://orcid.org/0009-0008-5623-2376","affiliations":[{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025399641","display_name":"Lukas Cavigelli","orcid":"https://orcid.org/0000-0003-1767-7715","observed_orcids":["https://orcid.org/0000-0003-1767-7715"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lukas Cavigelli","raw_affiliation_strings":["Computing Systems Lab, Huawei Zurich Research Center"],"raw_orcid":"https://orcid.org/0000-0003-1767-7715","affiliations":[{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148895635","display_name":"Luo Longfei","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Luo Longfei","raw_affiliation_strings":["Huawei Technologies Company Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Technologies Company Ltd","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148441643","display_name":"Jia Xinglei","orcid":"https://orcid.org/0009-0003-6160-7404","observed_orcids":["https://orcid.org/0009-0003-6160-7404"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jia Xinglei","raw_affiliation_strings":["Huawei Technologies Company Ltd"],"raw_orcid":"https://orcid.org/0009-0003-6160-7404","affiliations":[{"raw_affiliation_string":"Huawei Technologies Company Ltd","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050695684","display_name":"Onur Mutlu","orcid":"https://orcid.org/0000-0002-0075-2312","observed_orcids":["https://orcid.org/0000-0002-0075-2312"]},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Onur Mutlu","raw_affiliation_strings":["ETH Z\u00fcrich"],"raw_orcid":"https://orcid.org/0000-0002-0075-2312","affiliations":[{"raw_affiliation_string":"ETH Z\u00fcrich","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148436863","display_name":"Zhou Ke","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhou Ke","raw_affiliation_strings":["HUST"],"raw_orcid":"https://orcid.org/0000-0002-2161-8796","affiliations":[{"raw_affiliation_string":"HUST","institution_ids":["https://openalex.org/I47720641"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034711959","display_name":"Shai Aviram Bergman","orcid":"https://orcid.org/0000-0002-3484-1452","observed_orcids":["https://orcid.org/0000-0002-3484-1452"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shai Bergman","raw_affiliation_strings":["Computing Systems Lab, Huawei Zurich Research Center"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006337934","display_name":"Arash Tavakkol","orcid":"https://orcid.org/0000-0003-3859-1259","observed_orcids":["https://orcid.org/0000-0003-3859-1259"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Arash Tavakkol","raw_affiliation_strings":["Computing Systems Lab, Huawei Zurich Research Center"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Computing Systems Lab, Huawei Zurich Research Center","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5148876745","display_name":"Ji Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ji Zhang","raw_affiliation_strings":["Huawei Technologies Company Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Technologies Company Ltd","institution_ids":["https://openalex.org/I2250955327"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":2645,"currency":"USD","value_usd":2645},"apc_paid":{"value":2645,"currency":"USD","value_usd":2645},"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.65608656,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"25","issue":"2","first_page":"423","last_page":"427"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.6582000255584717,"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"}},"topics":[{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.6582000255584717,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.22679999470710754,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.014000000432133675,"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/flash-memory","display_name":"flash memory","score":0.7749999761581421},{"id":"https://openalex.org/keywords/llm-inference-acceleration","display_name":"LLM inference acceleration","score":0.7260000109672546},{"id":"https://openalex.org/keywords/memory-hierarchy","display_name":"memory hierarchy","score":0.35199999809265137},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.5619999766349792}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7946000099182129},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.6215000152587891},{"id":"https://openalex.org/C2777526259","wikidata":"https://www.wikidata.org/wiki/Q221836","display_name":"Flash (photography)","level":2,"score":0.5641999840736389},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.45170000195503235},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.27309998869895935},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2712000012397766},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.27070000767707825},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.25369998812675476},{"id":"https://openalex.org/C163258240","wikidata":"https://www.wikidata.org/wiki/Q25342","display_name":"Power (physics)","level":2,"score":0.25040000677108765},{"id":"https://openalex.org/C64260653","wikidata":"https://www.wikidata.org/wiki/Q1194864","display_name":"Electronic design automation","level":2,"score":0.23989999294281006}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/lca.2026.3729746","doi":"https://doi.org/10.1109/lca.2026.3729746","is_oa":true,"landing_page_url":"https://doi.org/10.1109/lca.2026.3729746","pdf_url":null,"source":{"id":"https://openalex.org/S17643076","display_name":"IEEE Computer Architecture Letters","issn_l":"1556-6056","issn":["1556-6056","1556-6064","2473-2575"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core","dongbi-b","norway-1","poland-70","russia-white-list-2"],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Computer Architecture Letters","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1109/lca.2026.3729746","doi":"https://doi.org/10.1109/lca.2026.3729746","is_oa":true,"landing_page_url":"https://doi.org/10.1109/lca.2026.3729746","pdf_url":null,"source":{"id":"https://openalex.org/S17643076","display_name":"IEEE Computer Architecture Letters","issn_l":"1556-6056","issn":["1556-6056","1556-6064","2473-2575"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core","dongbi-b","norway-1","poland-70","russia-white-list-2"],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Computer Architecture Letters","raw_type":"journal-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":11,"referenced_works":["https://openalex.org/W1544971428","https://openalex.org/W1998259214","https://openalex.org/W2010390358","https://openalex.org/W2014536047","https://openalex.org/W2158319074","https://openalex.org/W2949251802","https://openalex.org/W2962844766","https://openalex.org/W3036878841","https://openalex.org/W4390533494","https://openalex.org/W4409248462","https://openalex.org/W7127686171"],"related_works":["https://openalex.org/W7202268640","https://openalex.org/W7221503390","https://openalex.org/W7197043250","https://openalex.org/W7223250138","https://openalex.org/W7167784571","https://openalex.org/W7221061367","https://openalex.org/W7131867914","https://openalex.org/W7135093837","https://openalex.org/W7172432354","https://openalex.org/W7164931892"],"abstract_inverted_index":{"Large":[0],"language":[1,41],"model":[2,42],"(LLM)":[3,43],"inference":[4,187],"is":[5,20],"increasingly":[6],"constrained":[7],"by":[8,198,207],"accelerator":[9],"memory":[10,131],"capacity":[11,36],"rather":[12],"than":[13],"compute":[14],"throughput.":[15],"High":[16],"bandwidth":[17],"flash":[18,25,48,65,136,154],"(HBF)":[19],"an":[21,93,214,221],"emerging":[22],"3D-stacked":[23],"NAND":[24,64,135],"technology":[26],"that":[27,80,97,169,192],"provides":[28],"multi-terabyte":[29],"near-accelerator":[30],"capacity,":[31],"making":[32],"it":[33],"a":[34,85,109,122,166,226],"promising":[35],"tier":[37],"for":[38,113,162],"storing":[39],"large":[40],"weights.":[44],"However,":[45],"existing":[46,134],"high-bandwidth":[47],"(HBF)-based":[49],"proposals":[50],"face":[51],"three":[52,100,119],"adoption":[53],"challenges:":[54],"they":[55],"(i)":[56,121],"rely":[57],"on":[58],"coarse-grained":[59],"static":[60],"prefetching,":[61],"(ii)":[62,139],"expose":[63],"refresh":[66,144],"to":[67,84,91,174,213],"the":[68,146],"accelerator-visible":[69],"critical":[70,149],"path,":[71,150],"and":[72,151,183,201,203,210,225],"(iii)":[73,152],"assume":[74],"solid-state":[75],"drive":[76],"(SSD)-class":[77],"flash-management":[78],"mechanisms":[79],"are":[81],"poorly":[82],"matched":[83],"read-only":[86],"HBF":[87,95,111,127,176],"weight":[88,172],"tier.":[89],"Ourgoalis":[90],"design":[92],"efficient":[94],"substrate":[96,112],"addresses":[98],"these":[99],"challenges.":[101],"To":[102],"this":[103],"end,":[104],"we":[105],"propose":[106],"FLINT":[107,117,180,193],"(flashinferencetier),":[108],"workload-driven":[110],"capacity-scalable":[114],"LLM":[115,147,186,195],"inference.":[116],"introduces":[118],"mechanisms:":[120],"hardwareburst-buffer":[123],"controller,":[124],"which":[125,142,158],"pipelines":[126],"reads":[128],"with":[129,165],"accelerator-side":[130],"requests":[132],"using":[133],"page/cache":[137],"buffers,":[138],"aphantom-plane":[140],"refreshmechanism,":[141],"removes":[143],"from":[145],"decode":[148,196],"aread-only":[153],"translation":[155],"layer":[156],"(FTL),":[157],"replaces":[159],"SSD-class":[160],"support":[161],"arbitrary":[163],"writes":[164],"compact":[167],"table":[168],"translates":[170],"logical":[171],"bursts":[173],"physical":[175],"locations.":[177],"We":[178],"evaluate":[179],"across":[181],"dense":[182],"mixture-of-experts":[184],"(MoE)":[185],"models.":[188],"Our":[189],"evaluation":[190],"shows":[191],"improves":[194],"throughput":[197],"355\u00d7,":[199],"3.0\u00d7,":[200],"1.17\u00d7":[202],"reduces":[204],"energy":[205],"consumption":[206],"294\u00d7,":[208],"2.5\u00d7,":[209],"1.14\u00d7":[211],"compared":[212],"SSD-equipped":[215],"graphics":[216],"processing":[217],"unit":[218],"(GPU)":[219],"system,":[220,224,231],"HBM-only":[222],"GPU":[223,230],"prior":[227],"hybrid":[228],"HBM+HBF":[229],"respectively.":[232]},"counts_by_year":[],"updated_date":"2026-11-03T10:50:19.383937","created_date":"2026-09-02T00:00:00"}
