{"id":"https://openalex.org/W4415445242","doi":"https://doi.org/10.1109/tcad.2025.3624278","title":"MARCA-v2: Mamba Accelerator With Complementary State-Space Model Sparsity and Reconfigurable Architecture","display_name":"MARCA-v2: Mamba Accelerator With Complementary State-Space Model Sparsity and Reconfigurable Architecture","publication_year":2025,"publication_date":"2025-10-22","ids":{"openalex":"https://openalex.org/W4415445242","doi":"https://doi.org/10.1109/tcad.2025.3624278"},"language":null,"primary_location":{"id":"doi:10.1109/tcad.2025.3624278","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcad.2025.3624278","pdf_url":null,"source":{"id":"https://openalex.org/S100835903","display_name":"IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems","issn_l":"0278-0070","issn":["0278-0070","1937-4151"],"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 Transactions on Computer-Aided Design of Integrated Circuits and Systems","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/A5068894182","display_name":"Jinhao Li","orcid":"https://orcid.org/0009-0009-4286-6359"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jinhao Li","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0009-4286-6359","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109113870","display_name":"Shan Huang","orcid":"https://orcid.org/0009-0005-2012-8540"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shan Huang","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0005-2012-8540","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111878216","display_name":"Jianrong Xu","orcid":"https://orcid.org/0009-0000-7000-6537"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiaming Xu","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0000-7000-6537","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058051415","display_name":"J Liu","orcid":"https://orcid.org/0009-0003-8280-9072"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jun Liu","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0003-8280-9072","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018513988","display_name":"Ningyi Xu","orcid":"https://orcid.org/0009-0004-6809-7694"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ningyi Xu","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5102805465","display_name":"Guohao Dai","orcid":"https://orcid.org/0000-0002-8464-0130"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guohao Dai","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0000-0002-8464-0130","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I183067930"],"apc_list":null,"apc_paid":null,"fwci":1.377,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.83305876,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":"45","issue":"6","first_page":"2884","last_page":"2897"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10904","display_name":"Embedded Systems Design Techniques","score":0.9957000017166138,"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.9957000017166138,"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/T11044","display_name":"Particle Detector Development and Performance","score":0.9872999787330627,"subfield":{"id":"https://openalex.org/subfields/3106","display_name":"Nuclear and High Energy Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9764999747276306,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.7390000224113464},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.6610999703407288},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.4794999957084656},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.46230000257492065},{"id":"https://openalex.org/keywords/state","display_name":"State (computer science)","score":0.4487999975681305},{"id":"https://openalex.org/keywords/heuristic","display_name":"Heuristic","score":0.4449000060558319},{"id":"https://openalex.org/keywords/tree","display_name":"Tree (set theory)","score":0.3930000066757202},{"id":"https://openalex.org/keywords/state-space","display_name":"State space","score":0.3725999891757965}],"concepts":[{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.7390000224113464},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6664999723434448},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.6610999703407288},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.4794999957084656},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.46560001373291016},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.46230000257492065},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.4487999975681305},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.4449000060558319},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.40130001306533813},{"id":"https://openalex.org/C113174947","wikidata":"https://www.wikidata.org/wiki/Q2859736","display_name":"Tree (set theory)","level":2,"score":0.3930000066757202},{"id":"https://openalex.org/C72434380","wikidata":"https://www.wikidata.org/wiki/Q230930","display_name":"State space","level":2,"score":0.3725999891757965},{"id":"https://openalex.org/C158622935","wikidata":"https://www.wikidata.org/wiki/Q660848","display_name":"Nonlinear system","level":2,"score":0.37119999527931213},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.3366999924182892},{"id":"https://openalex.org/C129844170","wikidata":"https://www.wikidata.org/wiki/Q41299","display_name":"Quadratic equation","level":2,"score":0.3343999981880188},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.3199000060558319},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.3084999918937683},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.29679998755455017},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2896000146865845},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2831000089645386},{"id":"https://openalex.org/C123657996","wikidata":"https://www.wikidata.org/wiki/Q12271","display_name":"Architecture","level":2,"score":0.2824999988079071},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2775999903678894},{"id":"https://openalex.org/C162319229","wikidata":"https://www.wikidata.org/wiki/Q175263","display_name":"Data structure","level":2,"score":0.2773999869823456},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.2759000062942505},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.27309998869895935},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.26759999990463257},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.2660999894142151},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.2597000002861023}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tcad.2025.3624278","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcad.2025.3624278","pdf_url":null,"source":{"id":"https://openalex.org/S100835903","display_name":"IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems","issn_l":"0278-0070","issn":["0278-0070","1937-4151"],"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 Transactions on Computer-Aided Design of Integrated Circuits and Systems","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G6121003335","display_name":null,"funder_award_id":"24QB2706200","funder_id":"https://openalex.org/F4320327803","funder_display_name":"Shanghai Rising-Star Program"},{"id":"https://openalex.org/G6803541034","display_name":null,"funder_award_id":"U21B2031","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8262591375","display_name":null,"funder_award_id":"62104128","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320327803","display_name":"Shanghai Rising-Star Program","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0,70,192],"Language":[1],"models":[2],"with":[3,41,188,238,273,414,426,455,481],"state":[4,240],"space":[5,241],"model":[6,242,424],"(SSM)":[7,243],"especially":[8],"Mamba":[9,20,35,53,67,236,261,423],"have":[10],"demonstrated":[11],"remarkable":[12],"capabilities":[13],"in":[14,66,77,208,284,433],"various":[15,42],"domains.":[16],"Compared":[17],"to":[18,121,148,175,218,229,256,362,410],"Transformers,":[19],"reduces":[21],"the":[22,58,85,92,95,104,108,110,125,164,176,189,225,279,328,349,354,394,404,434,464],"quadratic":[23],"computational":[24],"complexity":[25],"and":[26,61,179,201,220,245,259,286,300,313,333,340,357,374,379,397,446,449,460,475,486],"achieves":[27,442,472],"a":[28,144,234,288,306,342],"higher":[29],"algorithm":[30,332],"performance.":[31],"Current":[32],"research":[33,49],"on":[34,38,50,80,132,252,327,388,422],"focuses":[36],"primarily":[37],"integrating":[39],"it":[40],"application":[43],"scenarios.":[44],"However,":[45],"there":[46],"is":[47,90],"limited":[48],"optimizing":[51],"for":[52,101,118,158,195,215,293,310,317],"processing.":[54],"Therefore,":[55],"we":[56,232,304,338],"profile":[57,278],"processing":[59],"carefully":[60],"identify":[62],"three":[63],"main":[64,205],"challenges":[65],"computations:":[68],"(1)":[69,269],"memory":[71,111],"access":[72,112],"overhead":[73,113,194],"of":[74,88,103,114,124,163,224,282],"element-wise":[75,115,130,151,319,324,401],"operations":[76,116,152,166,171],"SSM.":[78,209],"Based":[79,326],"<italic":[81,133,140,248,253,334,435,439,457,465,469,483],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[82,134,141,249,254,335,436,440,458,466,470,484],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">MARCA</i>":[83,135,142,336,459,485],"architecture,":[84,247,337],"time":[86],"proportion":[87],"SSM":[89,271,285,294],"still":[91,173],"bottleneck":[93],"when":[94],"sequence":[96],"length":[97],"reaches":[98],"2048,":[99],"accounting":[100],"62.52%":[102],"total":[105],"runtime.":[106],"Within":[107],"SSM,":[109],"account":[117],"97.17%,":[119],"leading":[120,217],"consuming":[122],"96.56%":[123],"time.":[126],"(2)":[127,321],"Inefficient":[128],"sparse":[129,156,159,302,318,323,375],"execution":[131,182],"architecture.":[136,325],"SOTA":[137],"architectures":[138],"like":[139],"propose":[143,233,287,305],"reconfigurable":[145,246,389,405],"reduction":[146],"tree":[147],"accelerate":[149],"dense":[150,373],"but":[153],"lack":[154],"effective":[155],"support":[157,257,371],"execution.":[160],"When":[161],"30%":[162],"elementwise":[165],"are":[167,172,203,266,407],"skipped,":[168],"these":[169,230],"skipped":[170],"mapped":[174],"PE":[177,364,390],"array":[178],"trigger":[180],"redundant":[181],"cycles,":[183],"incurring":[184],"1.43\u00d7":[185],"performance":[186],"gap":[187],"ideal.":[190],"(3)":[191,382],"area":[193,222],"nonlinear":[196,206,384,412],"function":[197,200,385,396],"unit.":[198],"Exponential":[199],"SiLU":[202,398],"two":[204,315],"functions":[207,413],"Previous":[210],"methods":[211],"design":[212,339],"specific":[213],"unit":[214,386],"acceleration,":[216],"38%":[219],"18%":[221],"overheads":[223],"PE.":[226],"In":[227,463],"response":[228],"challenges,":[231],"new":[235],"accelerator":[237],"complementary":[239,290],"sparsity":[244,272,292,331,355],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">MARCA-v2</i>,":[250],"based":[251,387],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">MARCA</i>,":[255],"fast":[258],"energy-efficient":[260],"computations.":[262],"Three":[263],"novel":[264],"techniques":[265],"as":[267],"follows:":[268],"Complementary":[270],"column-wise":[274,289],"granularity.":[275],"We":[276,392,418],"first":[277],"numerical":[280],"distributions":[281],"activations":[283],"static":[291],"computation.":[295],"To":[296],"further":[297],"enable":[298],"lightweight":[299,343],"hardware-friendly":[301],"computation,":[303],"\u03b4-bitmap":[307],"encoding":[308],"scheme":[309],"compressed":[311],"storage":[312],"introduce":[314],"abstractions":[316],"operations.":[320,402],"Lightweight":[322],"hardware":[329],"friendly":[330],"integrate":[341],"Metadata":[344],"Processing":[345],"Unit":[346],"(Meta-PU)":[347],"into":[348,399],"existing":[350],"pipeline,":[351],"which":[352],"decodes":[353],"metadata":[356],"dynamically":[358],"generates":[359],"control":[360],"signals":[361],"guide":[363],"arrays.":[365,391],"The":[366],"overall":[367],"architecture":[368],"can":[369],"efficiently":[370],"both":[372],"operations,":[376],"maximizing":[377],"speed":[378],"energy":[380,451,477],"efficiency.":[381],"Reusable":[383],"decompose":[393],"exponential":[395],"several":[400],"Thus,":[403],"PEs":[406],"fully":[408],"reused":[409],"execute":[411],"negligible":[415],"accuracy":[416],"loss.":[417],"conduct":[419],"extensive":[420],"experiments":[421],"families":[425],"different":[427],"sizes.":[428],"Experimental":[429],"results":[430],"show":[431],"that":[432],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">prefill</i>":[437],"stage,":[438,468],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">MARCA-v2</i>":[441,471],"1.77-10.87\u00d7,":[443],"1.03-":[444],"1.08\u00d7,":[445],"4.78-9.10\u00d7":[447],"speedup":[448,474],"8.29-33.47\u00d7/1.03-1.08\u00d7/4.78-9.10\u00d7":[450],"efficiency":[452,478],"improvement":[453,479],"compared":[454,480],"Mamba-GPU,":[456,482],"Spada,":[461,487],"respectively.":[462,488],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">decode</i>":[467],"0.88-7.65\u00d7/1.00-1.01\u00d7/1.19-1.64\u00d7":[473],"3.11-27.01\u00d7/1.00-1.01\u00d7/1.19-1.64\u00d7":[476]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-24T00:00:00"}