{"id":"https://openalex.org/W7165131753","doi":"https://doi.org/10.1145/3787109.3815254","title":"Mixed-precision Neural Networks on RISC-V CPU with Reconfigurable SIMD Instruction Extension via eFPGA","display_name":"Mixed-precision Neural Networks on RISC-V CPU with Reconfigurable SIMD Instruction Extension via eFPGA","publication_year":2026,"publication_date":"2026-06-18","ids":{"openalex":"https://openalex.org/W7165131753","doi":"https://doi.org/10.1145/3787109.3815254"},"language":null,"primary_location":{"id":"doi:10.1145/3787109.3815254","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3787109.3815254","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Great Lakes Symposium on VLSI 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3787109.3815254","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5086043351","display_name":"Zixin Yang","orcid":"https://orcid.org/0000-0002-3754-8542"},"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":"Zixin Yang","raw_affiliation_strings":["State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0003-1797-0118","affiliations":[{"raw_affiliation_string":"State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102014346","display_name":"Yanze Li","orcid":"https://orcid.org/0000-0003-4647-4482"},"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":"Yanze Li","raw_affiliation_strings":["State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0006-6131-9199","affiliations":[{"raw_affiliation_string":"State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101935854","display_name":"Zhichao Wei","orcid":"https://orcid.org/0000-0002-4796-9690"},"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":"Zhichao Wei","raw_affiliation_strings":["State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0004-4170-9320","affiliations":[{"raw_affiliation_string":"State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103809501","display_name":"Jian Wang","orcid":"https://orcid.org/0000-0002-6086-1924"},"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":"Jian Wang","raw_affiliation_strings":["State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China"],"raw_orcid":"https://orcid.org/0000-0002-6086-1924","affiliations":[{"raw_affiliation_string":"State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5081419061","display_name":"Jinmei Lai","orcid":"https://orcid.org/0009-0003-5238-4720"},"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":"Jinmei Lai","raw_affiliation_strings":["State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0003-5238-4720","affiliations":[{"raw_affiliation_string":"State Key Laboratory of Integrated Chip & System, Fudan university, Shanghai, China","institution_ids":["https://openalex.org/I24943067"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I24943067"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.76843223,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"426","last_page":"433"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.4934000074863434,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.4934000074863434,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10320","display_name":"Neural Networks and Applications","score":0.0771000012755394,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.04610000178217888,"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/simd","display_name":"SIMD","score":0.8098999857902527},{"id":"https://openalex.org/keywords/memory-bandwidth","display_name":"Memory bandwidth","score":0.45159998536109924},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.428600013256073},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.41339999437332153},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.3813999891281128},{"id":"https://openalex.org/keywords/adaptability","display_name":"Adaptability","score":0.3716999888420105},{"id":"https://openalex.org/keywords/flexibility","display_name":"Flexibility (engineering)","score":0.34060001373291016},{"id":"https://openalex.org/keywords/bandwidth","display_name":"Bandwidth (computing)","score":0.33230000734329224}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8561999797821045},{"id":"https://openalex.org/C150552126","wikidata":"https://www.wikidata.org/wiki/Q339387","display_name":"SIMD","level":2,"score":0.8098999857902527},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.5680999755859375},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5396999716758728},{"id":"https://openalex.org/C188045654","wikidata":"https://www.wikidata.org/wiki/Q17148339","display_name":"Memory bandwidth","level":2,"score":0.45159998536109924},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.428600013256073},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.41339999437332153},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.3813999891281128},{"id":"https://openalex.org/C177606310","wikidata":"https://www.wikidata.org/wiki/Q5674297","display_name":"Adaptability","level":2,"score":0.3716999888420105},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.3598000109195709},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.34060001373291016},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.33230000734329224},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.32089999318122864},{"id":"https://openalex.org/C49154492","wikidata":"https://www.wikidata.org/wiki/Q5300","display_name":"Central processing unit","level":2,"score":0.31529998779296875},{"id":"https://openalex.org/C2778100165","wikidata":"https://www.wikidata.org/wiki/Q1589327","display_name":"Memory hierarchy","level":3,"score":0.3093000054359436},{"id":"https://openalex.org/C47487241","wikidata":"https://www.wikidata.org/wiki/Q5227230","display_name":"Data access","level":2,"score":0.3034000098705292},{"id":"https://openalex.org/C202491316","wikidata":"https://www.wikidata.org/wiki/Q272683","display_name":"Instruction set","level":2,"score":0.28870001435279846},{"id":"https://openalex.org/C26771161","wikidata":"https://www.wikidata.org/wiki/Q16980","display_name":"ARM architecture","level":2,"score":0.2777999937534332},{"id":"https://openalex.org/C2779808786","wikidata":"https://www.wikidata.org/wiki/Q6664603","display_name":"Locality","level":2,"score":0.2777999937534332},{"id":"https://openalex.org/C199833920","wikidata":"https://www.wikidata.org/wiki/Q612536","display_name":"Vector quantization","level":2,"score":0.27239999175071716},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.2621999979019165},{"id":"https://openalex.org/C161824985","wikidata":"https://www.wikidata.org/wiki/Q919509","display_name":"Vector processor","level":2,"score":0.25839999318122864}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3787109.3815254","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3787109.3815254","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Great Lakes Symposium on VLSI 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3787109.3815254","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3787109.3815254","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Great Lakes Symposium on VLSI 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","display_name":"Affordable and clean energy","score":0.8775021433830261}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":9,"referenced_works":["https://openalex.org/W3047342117","https://openalex.org/W3129912126","https://openalex.org/W3160646870","https://openalex.org/W3205109693","https://openalex.org/W4312298069","https://openalex.org/W4360831846","https://openalex.org/W4391547488","https://openalex.org/W4402426679","https://openalex.org/W4414354696"],"related_works":[],"abstract_inverted_index":{"Mixed-precision":[0],"quantization":[1],"approach":[2],"has":[3],"become":[4],"a":[5,74,97,152],"key":[6],"technique":[7],"for":[8,26,80,102],"deploying":[9],"neural":[10,145],"networks":[11],"(NNs)":[12],"on":[13,121,151,169],"Central":[14],"Processing":[15],"Units":[16],"(CPUs).":[17],"However,":[18],"current":[19],"CPUs":[20],"face":[21],"the":[22,42,91,106,116,141,170,189],"following":[23],"major":[24],"limitations":[25],"mixed-precision":[27,36,86,174],"NNs:":[28],"Instruction":[29],"Set":[30],"Architecture":[31],"(ISA)":[32],"lack":[33],"adaptability":[34],"to":[35,46,57,133],"SIMD":[37,82,112,137],"instructions,":[38],"bandwidth":[39,136],"bottlenecks":[40],"in":[41,105],"memory":[43],"system":[44],"due":[45],"large":[47],"amounts":[48],"of":[49,143],"vector":[50,92,127],"data,":[51],"insufficient":[52],"programmability":[53],"and":[54,109,129,159,183],"flexibility":[55],"lead":[56],"fragile":[58],"over-optimization":[59],"as":[60],"applications":[61],"change":[62],"rapidly.":[63],"In":[64],"this":[65],"work,":[66],"we":[67],"demonstrate":[68,164],"an":[69,179],"extended":[70],"RISC-V":[71,117],"CPU":[72,107],"with":[73,96,125,172],"customized":[75],"tightly-coupled":[76],"embedded":[77],"FPGA":[78],"(eFPGA)":[79],"reconfigurable":[81],"instructions":[83,113],"extension,":[84],"targeting":[85],"NNs":[87],"deployment.":[88],"We":[89,119,139],"optimize":[90],"data":[93,100],"mapping":[94],"strategy":[95],"configurable":[98],"precision":[99],"path":[101],"eFPGA":[103,154,171],"access":[104],"extension":[108],"design":[110,131],"innovative":[111],"that":[114,165],"extend":[115],"ISA.":[118],"focus":[120],"cache":[122],"hierarchy":[123],"optimization":[124],"related":[126],"register":[128],"LSU":[130],"choices":[132],"achieve":[134,178],"high":[135],"operations.":[138],"enable":[140],"implementation":[142],"high-throughput":[144],"MAC":[146],"operations":[147],"at":[148],"different":[149],"precisions":[150],"resource-constrained":[153],"fabric":[155],"via":[156],"unpacking":[157],"units":[158],"lane-based":[160],"parallelism.":[161],"Experimental":[162],"results":[163],"our":[166],"approach,":[167],"performed":[168],"representative":[173],"quantized":[175],"NNs,":[176],"can":[177],"average":[180],"23.7\u00d7":[181],"performance":[182],"9.2\u00d7":[184],"energy":[185],"efficiency":[186],"gains":[187],"over":[188],"baseline":[190],"processor.":[191]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-06-19T00:00:00"}