{"id":"https://openalex.org/W7131425448","doi":"https://doi.org/10.1109/mm.2026.3667115","title":"Demystifying the Cost Versus Benefits of Sparse Large Language Model Acceleration","display_name":"Demystifying the Cost Versus Benefits of Sparse Large Language Model Acceleration","publication_year":2026,"publication_date":"2026-02-25","ids":{"openalex":"https://openalex.org/W7131425448","doi":"https://doi.org/10.1109/mm.2026.3667115"},"language":null,"primary_location":{"id":"doi:10.1109/mm.2026.3667115","is_oa":false,"landing_page_url":"https://doi.org/10.1109/mm.2026.3667115","pdf_url":null,"source":{"id":"https://openalex.org/S59697426","display_name":"IEEE Micro","issn_l":"0272-1732","issn":["0272-1732","1937-4143"],"is_oa":false,"is_in_doaj":false,"is_core":true,"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":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Micro","raw_type":"journal-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/A5008320997","display_name":"Abhimanyu Rajeshkumar Bambhaniya","orcid":"https://orcid.org/0000-0003-0145-9713"},"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":"Abhimanyu R. Bambhaniya","raw_affiliation_strings":["Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA"],"raw_orcid":"https://orcid.org/0000-0003-0145-9713","affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA","institution_ids":["https://openalex.org/I130701444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5035130804","display_name":"Sheng-Chun Kao","orcid":"https://orcid.org/0000-0001-7928-9027"},"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":"Sheng-Chun Kao","raw_affiliation_strings":["Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA"],"raw_orcid":"https://orcid.org/0000-0001-7928-9027","affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA","institution_ids":["https://openalex.org/I130701444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069792279","display_name":"Geonhwa Jeong","orcid":"https://orcid.org/0000-0001-6659-3927"},"institutions":[{"id":"https://openalex.org/I4210099336","display_name":"Menlo School","ror":"https://ror.org/01240pn49","country_code":"US","type":"education","lineage":["https://openalex.org/I4210099336"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Geonhwa Jeong","raw_affiliation_strings":["Meta, Menlo Park, CA, USA"],"raw_orcid":"https://orcid.org/0000-0001-6659-3927","affiliations":[{"raw_affiliation_string":"Meta, Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210099336"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126792685","display_name":"Suvinay Subrimaniam","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Suvinay Subramanian","raw_affiliation_strings":["Google, Mountain View, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-8715-8964","affiliations":[{"raw_affiliation_string":"Google, Mountain View, CA, USA","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110993207","display_name":"Amir Yazdanbaksh","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Amir Yazdanbakhsh","raw_affiliation_strings":["Google DeepMind, Mountain View, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-8531-1873","affiliations":[{"raw_affiliation_string":"Google DeepMind, Mountain View, CA, USA","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5034089074","display_name":"Tushar Krishna","orcid":"https://orcid.org/0000-0001-5738-6942"},"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":"Tushar Krishna","raw_affiliation_strings":["School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA"],"raw_orcid":"https://orcid.org/0000-0001-5738-6942","affiliations":[{"raw_affiliation_string":"School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA","institution_ids":["https://openalex.org/I130701444"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"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":0.17161208,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"46","issue":"2","first_page":"38","last_page":"50"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10559","display_name":"Particle Accelerators and Free-Electron Lasers","score":0.16509999334812164,"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/T10559","display_name":"Particle Accelerators and Free-Electron Lasers","score":0.16509999334812164,"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/T10857","display_name":"Advanced Electron Microscopy Techniques and Applications","score":0.06639999896287918,"subfield":{"id":"https://openalex.org/subfields/1315","display_name":"Structural Biology"},"field":{"id":"https://openalex.org/fields/13","display_name":"Biochemistry, Genetics and Molecular Biology"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T11367","display_name":"Particle accelerators and beam dynamics","score":0.04450000077486038,"subfield":{"id":"https://openalex.org/subfields/2202","display_name":"Aerospace 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/exploit","display_name":"Exploit","score":0.7243000268936157},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.7111999988555908},{"id":"https://openalex.org/keywords/design-space-exploration","display_name":"Design space exploration","score":0.5558000206947327},{"id":"https://openalex.org/keywords/range","display_name":"Range (aeronautics)","score":0.49300000071525574},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.4156999886035919},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.414900004863739},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.41429999470710754},{"id":"https://openalex.org/keywords/interconnection","display_name":"Interconnection","score":0.3684999942779541}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8693000078201294},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.7243000268936157},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.7111999988555908},{"id":"https://openalex.org/C2776221188","wikidata":"https://www.wikidata.org/wiki/Q21072556","display_name":"Design space exploration","level":2,"score":0.5558000206947327},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.519599974155426},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.49300000071525574},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.4456999897956848},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.4156999886035919},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.414900004863739},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.41429999470710754},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.3935999870300293},{"id":"https://openalex.org/C123745756","wikidata":"https://www.wikidata.org/wiki/Q1665949","display_name":"Interconnection","level":2,"score":0.3684999942779541},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.352400004863739},{"id":"https://openalex.org/C2776834041","wikidata":"https://www.wikidata.org/wiki/Q25346349","display_name":"Execution model","level":2,"score":0.34850001335144043},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.3483000099658966},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3393999934196472},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.30090001225471497},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.29319998621940613},{"id":"https://openalex.org/C60832428","wikidata":"https://www.wikidata.org/wiki/Q945818","display_name":"Partitioned global address space","level":3,"score":0.27239999175071716},{"id":"https://openalex.org/C58013763","wikidata":"https://www.wikidata.org/wiki/Q5754574","display_name":"High-level synthesis","level":3,"score":0.26969999074935913},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.2669000029563904},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.26170000433921814},{"id":"https://openalex.org/C138852830","wikidata":"https://www.wikidata.org/wiki/Q2292993","display_name":"Design methods","level":2,"score":0.2554999887943268},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.2554999887943268}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/mm.2026.3667115","is_oa":false,"landing_page_url":"https://doi.org/10.1109/mm.2026.3667115","pdf_url":null,"source":{"id":"https://openalex.org/S59697426","display_name":"IEEE Micro","issn_l":"0272-1732","issn":["0272-1732","1937-4143"],"is_oa":false,"is_in_doaj":false,"is_core":true,"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":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Micro","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":12,"referenced_works":["https://openalex.org/W1985818188","https://openalex.org/W2604670770","https://openalex.org/W2725159389","https://openalex.org/W3016542674","https://openalex.org/W3040646053","https://openalex.org/W3190062760","https://openalex.org/W4285335127","https://openalex.org/W4308083753","https://openalex.org/W4360831840","https://openalex.org/W4385245566","https://openalex.org/W4406164090","https://openalex.org/W4412042418"],"related_works":[],"abstract_inverted_index":{"The":[0],"rapid":[1],"growth":[2],"in":[3,40,118],"LLM":[4,142],"model":[5,77],"sizes":[6],"has":[7],"intensified":[8],"the":[9,103,129],"need":[10],"for":[11,66,139],"sparsification":[12],"techniques":[13],"to":[14,25,33,72],"reduce":[15],"computational":[16],"and":[17,52,76,89,113,121,135],"memory":[18],"overheads.":[19],"This":[20],"presents":[21],"a":[22,57,95],"unique":[23],"opportunity":[24],"design":[26,116,130],"accelerator":[27],"chips":[28],"that":[29,62],"can":[30],"exploit":[31],"sparsity":[32,74,108],"improve":[34],"runtime":[35],"efficiency.":[36,125],"Supporting":[37],"sparse":[38,69,133],"execution":[39],"AI":[41,99],"accelerators":[42,70,134],"requires":[43],"coordinated":[44],"enhancements":[45],"across":[46,94],"three":[47],"key":[48,110],"subsystems:":[49],"compute,":[50,119],"memory,":[51,120],"interconnect.":[53],"However,":[54],"there":[55],"is":[56],"lack":[58],"of":[59,98,106,132],"systematic":[60],"methodologies":[61],"provide":[63],"architectural":[64,104],"insights":[65],"building":[67],"efficient":[68],"tailored":[71],"diverse":[73],"patterns":[75],"architectures.":[78],"In":[79],"this":[80],"work,":[81],"we":[82,127],"systematically":[83],"analyze":[84],"trade-offs":[85],"between":[86],"area,":[87],"energy,":[88],"performance":[90],"under":[91,144],"quality":[92],"constraints":[93],"broad":[96],"range":[97],"models.":[100],"We":[101],"quantify":[102],"impact":[105],"enabling":[107],"on":[109],"hardware":[111],"subsystems,":[112],"characterize":[114],"how":[115],"choices":[117],"interconnect":[122],"affect":[123],"overall":[124],"Finally,":[126],"explore":[128],"space":[131],"identify":[136],"optimal":[137],"configurations":[138],"accelerating":[140],"modern":[141],"workloads":[143],"practical":[145],"constraints.":[146]},"counts_by_year":[],"updated_date":"2026-05-04T06:00:05.808912","created_date":"2026-02-26T00:00:00"}
