{"id":"https://openalex.org/W7106151227","doi":"https://doi.org/10.1109/iccad66269.2025.11240845","title":"Robin: RWKV Accelerator using Block Circulant Matrices based on FPGA","display_name":"Robin: RWKV Accelerator using Block Circulant Matrices based on FPGA","publication_year":2025,"publication_date":"2025-10-26","ids":{"openalex":"https://openalex.org/W7106151227","doi":"https://doi.org/10.1109/iccad66269.2025.11240845"},"language":null,"primary_location":{"id":"doi:10.1109/iccad66269.2025.11240845","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccad66269.2025.11240845","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/ACM International Conference On Computer Aided Design (ICCAD)","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":null,"display_name":"Zeyu Li","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Zeyu Li","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Shangkun Li","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Shangkun Li","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Chuyi Dai","orcid":null},"institutions":[{"id":"https://openalex.org/I24943067","display_name":"Fudan University","ror":"https://ror.org/013q1eq08","country_code":"CN","type":"education","lineage":["https://openalex.org/I24943067"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chuyi Dai","raw_affiliation_strings":["Fudan University"],"affiliations":[{"raw_affiliation_string":"Fudan University","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Chaofang Ma","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Chaofang Ma","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Jiawei Liang","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Jiawei Liang","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Xin Li","orcid":null},"institutions":[{"id":"https://openalex.org/I4210159968","display_name":"Duke Kunshan University","ror":"https://ror.org/04sr5ys16","country_code":"CN","type":"education","lineage":["https://openalex.org/I170897317","https://openalex.org/I37461747","https://openalex.org/I4210159968"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xin Li","raw_affiliation_strings":["Duke Kunshan University"],"affiliations":[{"raw_affiliation_string":"Duke Kunshan University","institution_ids":["https://openalex.org/I4210159968"]}]},{"author_position":"last","author":{"id":null,"display_name":"Wei Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Wei Zhang","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":7,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I200769079"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.60932798,"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":"9"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10904","display_name":"Embedded Systems Design Techniques","score":0.2037000060081482,"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":0.2037000060081482,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.17440000176429749,"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/T11697","display_name":"Numerical Methods and Algorithms","score":0.17080000042915344,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.7440999746322632},{"id":"https://openalex.org/keywords/workflow","display_name":"Workflow","score":0.4636000096797943},{"id":"https://openalex.org/keywords/reconfigurable-computing","display_name":"Reconfigurable computing","score":0.45890000462532043},{"id":"https://openalex.org/keywords/digital-signal-processing","display_name":"Digital signal processing","score":0.4275999963283539},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.41670000553131104},{"id":"https://openalex.org/keywords/efficient-energy-use","display_name":"Efficient energy use","score":0.400299996137619},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.38920000195503235},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.3889000117778778},{"id":"https://openalex.org/keywords/matrix-multiplication","display_name":"Matrix multiplication","score":0.382099986076355},{"id":"https://openalex.org/keywords/cuda","display_name":"CUDA","score":0.3792000114917755}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7800999879837036},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.7440999746322632},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.47429999709129333},{"id":"https://openalex.org/C177212765","wikidata":"https://www.wikidata.org/wiki/Q627335","display_name":"Workflow","level":2,"score":0.4636000096797943},{"id":"https://openalex.org/C142962650","wikidata":"https://www.wikidata.org/wiki/Q240838","display_name":"Reconfigurable computing","level":3,"score":0.45890000462532043},{"id":"https://openalex.org/C84462506","wikidata":"https://www.wikidata.org/wiki/Q173142","display_name":"Digital signal processing","level":2,"score":0.4275999963283539},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.41670000553131104},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.40610000491142273},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.400299996137619},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.38920000195503235},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.3889000117778778},{"id":"https://openalex.org/C17349429","wikidata":"https://www.wikidata.org/wiki/Q1049914","display_name":"Matrix multiplication","level":3,"score":0.382099986076355},{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.3792000114917755},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.37549999356269836},{"id":"https://openalex.org/C115973184","wikidata":"https://www.wikidata.org/wiki/Q245457","display_name":"Circulant matrix","level":2,"score":0.3481999933719635},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3400999903678894},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.33059999346733093},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.32670000195503235},{"id":"https://openalex.org/C63000827","wikidata":"https://www.wikidata.org/wiki/Q3080428","display_name":"Software portability","level":2,"score":0.3165000081062317},{"id":"https://openalex.org/C174348530","wikidata":"https://www.wikidata.org/wiki/Q188635","display_name":"Bridging (networking)","level":2,"score":0.31470000743865967},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.3142000138759613},{"id":"https://openalex.org/C58013763","wikidata":"https://www.wikidata.org/wiki/Q5754574","display_name":"High-level synthesis","level":3,"score":0.3107999861240387},{"id":"https://openalex.org/C164620267","wikidata":"https://www.wikidata.org/wiki/Q376953","display_name":"Adder","level":3,"score":0.3028999865055084},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.296999990940094},{"id":"https://openalex.org/C79581498","wikidata":"https://www.wikidata.org/wiki/Q1367530","display_name":"Suite","level":2,"score":0.2946999967098236},{"id":"https://openalex.org/C2778325283","wikidata":"https://www.wikidata.org/wiki/Q1125244","display_name":"Logic block","level":3,"score":0.27730000019073486},{"id":"https://openalex.org/C127964446","wikidata":"https://www.wikidata.org/wiki/Q1092142","display_name":"Computational resource","level":3,"score":0.2766000032424927},{"id":"https://openalex.org/C71901391","wikidata":"https://www.wikidata.org/wiki/Q7126699","display_name":"Upload","level":2,"score":0.27570000290870667},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.271699994802475},{"id":"https://openalex.org/C172430144","wikidata":"https://www.wikidata.org/wiki/Q17111997","display_name":"Symmetric multiprocessor system","level":2,"score":0.2676999866962433},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.26660001277923584},{"id":"https://openalex.org/C149810388","wikidata":"https://www.wikidata.org/wiki/Q5374873","display_name":"Emulation","level":2,"score":0.2662000060081482},{"id":"https://openalex.org/C86111242","wikidata":"https://www.wikidata.org/wiki/Q859595","display_name":"Coprocessor","level":2,"score":0.2630000114440918},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.26030001044273376},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.257099986076355},{"id":"https://openalex.org/C66429209","wikidata":"https://www.wikidata.org/wiki/Q7248499","display_name":"Programming complexity","level":5,"score":0.2538999915122986},{"id":"https://openalex.org/C2776277307","wikidata":"https://www.wikidata.org/wiki/Q22074755","display_name":"Stratix","level":3,"score":0.2517000138759613}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iccad66269.2025.11240845","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccad66269.2025.11240845","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/ACM International Conference On Computer Aided Design (ICCAD)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7","score":0.8986954689025879}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":17,"referenced_works":["https://openalex.org/W2890894339","https://openalex.org/W2946609015","https://openalex.org/W2962820060","https://openalex.org/W2963015836","https://openalex.org/W2970062726","https://openalex.org/W2998617917","https://openalex.org/W3047848469","https://openalex.org/W3138516171","https://openalex.org/W3194676777","https://openalex.org/W3214431650","https://openalex.org/W4385245566","https://openalex.org/W4389524555","https://openalex.org/W4393578753","https://openalex.org/W4401211806","https://openalex.org/W4402670692","https://openalex.org/W4409369724","https://openalex.org/W4415797212"],"related_works":[],"abstract_inverted_index":{"Recent":[0],"advancements":[1],"in":[2,213,219],"linear-attention":[3],"models,":[4],"such":[5],"as":[6],"RWKV,":[7],"have":[8],"opened":[9],"up":[10,208],"new":[11],"possibilities":[12],"for":[13,34,91,234],"efficient":[14],"sequence":[15],"processing":[16],"by":[17,38,82],"reducing":[18],"the":[19,95,101,116,180],"computational":[20,45,107],"overhead":[21],"of":[22,97,183],"traditional":[23],"Transformer":[24],"architectures.":[25],"Field-programmable":[26],"gate":[27],"arrays":[28],"(FPGAs)":[29],"offer":[30],"a":[31,133,147,172,210,216,231],"compelling":[32,232],"solution":[33,233],"deep":[35],"learning":[36,68],"applications":[37],"providing":[39],"customizable":[40],"hardware":[41,76],"architectures":[42],"that":[43,138,177],"enhance":[44],"efficiency":[46,182,221],"and":[47,75,85,121,142,195,215],"flexibility.":[48],"However,":[49],"deploying":[50,235],"these":[51,128],"models":[52,237],"on":[53,65,100,152,167,238],"FPGAs":[54,79,120],"introduces":[55],"several":[56],"challenges.":[57],"Previous":[58],"FPGA":[59,135],"deployment":[60],"workflows":[61],"tend":[62],"to":[63,105,114,191,206,209,223],"focus":[64],"general":[66],"machine":[67],"tasks,":[69],"lacking":[70],"sufficient":[71],"integration":[72],"between":[73,119],"software":[74],"optimizations.":[77],"Besides,":[78],"are":[80],"constrained":[81],"limited":[83],"on-chip":[84],"off-chip":[86],"memory,":[87],"posing":[88],"significant":[89,106],"challenges":[90],"weight":[92,148],"storage.":[93],"Moreover,":[94],"predominance":[96],"linear":[98],"operations":[99],"GPU":[102,227],"runtime":[103],"leads":[104],"bottlenecks.":[108],"These":[109],"obstacles":[110],"necessitate":[111],"innovative":[112],"solutions":[113],"bridge":[115],"performance":[117],"gap":[118],"GPUs":[122],"while":[123,163],"preserving":[124],"model":[125],"accuracyTo":[126],"overcome":[127],"challenges,":[129],"we":[130],"introduce":[131],"Robin,":[132],"fine-grained":[134],"accelerator":[136],"workflow":[137],"integrates":[139],"both":[140,193],"algorithm-level":[141],"hardware-level":[143],"optimization.":[144],"Robin":[145,205],"leverages":[146],"compression":[149],"technique":[150],"based":[151],"Partial":[153],"Block":[154],"Circulant":[155],"Matrices":[156],"(PBCM),":[157],"which":[158],"effectively":[159],"reduces":[160],"storage":[161],"demands":[162],"maintaining":[164],"accuracy.":[165],"Based":[166],"PBCM,":[168],"our":[169],"design":[170],"employs":[171],"configurable":[173],"circulant":[174,194],"computing":[175],"core":[176],"fully":[178],"exploits":[179],"bit-width":[181],"DSP48E":[184],"resources":[185],"through":[186],"two":[187],"DSP":[188],"packaging":[189],"strategies":[190],"support":[192],"standard":[196],"matrix":[197],"operations.":[198],"The":[199],"combined":[200],"end-to-end":[201],"software-hardware":[202],"co-design":[203],"enables":[204],"achieve":[207],"3.09\u00d7":[211],"increase":[212],"throughput":[214],"7.31\u00d7":[217],"boost":[218],"energy":[220],"compared":[222],"high-end":[224],"Tesla":[225],"A100":[226],"implementations,":[228],"making":[229],"it":[230],"RWKV":[236],"FPGAs.":[239]},"counts_by_year":[],"updated_date":"2026-01-25T23:04:38.658462","created_date":"2025-11-20T00:00:00"}