{"id":"https://openalex.org/W7134943445","doi":"https://doi.org/10.1109/asp-dac66049.2026.11420284","title":"dLLM-OPU: An FPGA Overlay Processor for Accelerated Diffusion Large Language Models","display_name":"dLLM-OPU: An FPGA Overlay Processor for Accelerated Diffusion Large Language Models","publication_year":2026,"publication_date":"2026-01-19","ids":{"openalex":"https://openalex.org/W7134943445","doi":"https://doi.org/10.1109/asp-dac66049.2026.11420284"},"language":null,"primary_location":{"id":"doi:10.1109/asp-dac66049.2026.11420284","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asp-dac66049.2026.11420284","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 31st Asia and South Pacific Design Automation Conference (ASP-DAC)","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":"https://openalex.org/A5117428653","display_name":"Yangbo Wei","orcid":null},"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":"Yangbo Wei","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":"middle","author":{"id":"https://openalex.org/A5101262539","display_name":"Shaoqiang Lu","orcid":null},"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":"Shaoqiang Lu","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":"middle","author":{"id":"https://openalex.org/A5004536230","display_name":"Junhong Qian","orcid":null},"institutions":[{"id":"https://openalex.org/I76569877","display_name":"Southeast University","ror":"https://ror.org/04ct4d772","country_code":"CN","type":"education","lineage":["https://openalex.org/I76569877"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Junhong Qian","raw_affiliation_strings":["Southeast University,Nanjing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Southeast University,Nanjing,China","institution_ids":["https://openalex.org/I76569877"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128752832","display_name":"Lei He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lei He","raw_affiliation_strings":["Eastern Institute of Technology,Ningbo,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Eastern Institute of Technology,Ningbo,China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5049051725","display_name":"Dongge Qin","orcid":null},"institutions":[{"id":"https://openalex.org/I76569877","display_name":"Southeast University","ror":"https://ror.org/04ct4d772","country_code":"CN","type":"education","lineage":["https://openalex.org/I76569877"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dongge Qin","raw_affiliation_strings":["Southeast University,Nanjing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Southeast University,Nanjing,China","institution_ids":["https://openalex.org/I76569877"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128699136","display_name":"Xiao Shi","orcid":null},"institutions":[{"id":"https://openalex.org/I76569877","display_name":"Southeast University","ror":"https://ror.org/04ct4d772","country_code":"CN","type":"education","lineage":["https://openalex.org/I76569877"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiao Shi","raw_affiliation_strings":["Southeast University,Nanjing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Southeast University,Nanjing,China","institution_ids":["https://openalex.org/I76569877"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128712754","display_name":"Chen Wu","orcid":null},"institutions":[{"id":"https://openalex.org/I4210165339","display_name":"Ningbo Institute of Industrial Technology","ror":"https://ror.org/05nqg3g04","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210165339"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chen Wu","raw_affiliation_strings":["Ningbo Institute of Digital Twin Eastern Institute of Technology,Ningbo,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Ningbo Institute of Digital Twin Eastern Institute of Technology,Ningbo,China","institution_ids":["https://openalex.org/I4210165339"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5128719480","display_name":"Linfeng Zhang","orcid":null},"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":"Linfeng Zhang","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"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":8,"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":0.28967004,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1476","last_page":"1482"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.1793999969959259,"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":0.1793999969959259,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.10419999808073044,"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":0.1006999984383583,"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/overlay","display_name":"Overlay","score":0.6075000166893005},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.5088000297546387},{"id":"https://openalex.org/keywords/diffusion","display_name":"Diffusion","score":0.3691999912261963},{"id":"https://openalex.org/keywords/coupling","display_name":"Coupling (piping)","score":0.27489998936653137},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.2687999904155731}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6735000014305115},{"id":"https://openalex.org/C136085584","wikidata":"https://www.wikidata.org/wiki/Q910289","display_name":"Overlay","level":2,"score":0.6075000166893005},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.5088000297546387},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.3691999912261963},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.34769999980926514},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.3111000061035156},{"id":"https://openalex.org/C131584629","wikidata":"https://www.wikidata.org/wiki/Q4308705","display_name":"Coupling (piping)","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.26910001039505005},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2687999904155731},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2669999897480011}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/asp-dac66049.2026.11420284","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asp-dac66049.2026.11420284","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2026 31st Asia and South Pacific Design Automation Conference (ASP-DAC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":4,"referenced_works":["https://openalex.org/W2979455536","https://openalex.org/W4402778510","https://openalex.org/W4408151450","https://openalex.org/W4409285540"],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"are":[4],"achieving":[5],"unprecedented":[6],"performance":[7],"across":[8],"diverse":[9],"tasks,":[10],"benefiting":[11],"from":[12,65],"autoregressive":[13],"generation.":[14],"However,":[15,47],"this":[16],"left-to-right":[17],"decoding":[18,168],"paradigm":[19],"inherently":[20],"limits":[21],"contextual":[22],"understanding":[23],"quality.":[24,46],"Diffusion-based":[25],"LLMs":[26],"(dLLMs)":[27],"offer":[28],"a":[29,91,111,133,165,172],"promising":[30],"alternative":[31],"by":[32,145,153],"iteratively":[33],"refining":[34],"sequences":[35],"via":[36,149],"denoising,":[37,60],"enabling":[38],"stronger":[39],"bidirectional":[40],"context":[41],"modeling":[42],"and":[43,55,61,109,164,181],"improved":[44],"generation":[45],"dLLMs":[48],"face":[49],"two":[50,87],"main":[51],"challenges:":[52],"redundant":[53],"computation":[54],"memory":[56,161],"overhead":[57],"in":[58,189],"multi-step":[59],"excessive":[62],"inference":[63],"cost":[64],"over-denoising":[66],"under":[67],"fixed-step":[68],"schedules.":[69],"To":[70],"address":[71],"these":[72,130],"issues,":[73],"we":[74],"propose":[75],"dLLM-OPU,":[76],"an":[77,154],"FPGA":[78],"overlay":[79],"processor":[80],"to":[81],"accelerate":[82],"dLLMs.":[83],"Our":[84],"solution":[85],"features":[86],"key":[88],"innovations:":[89],"(1)":[90],"Region-Adaptive":[92],"Caching":[93],"for":[94,103],"Dynamic":[95],"Column":[96],"Sparsity":[97],"Framework":[98],"that":[99,117,140,159],"exploits":[100],"temporal":[101],"locality":[102],"selective":[104],"recomputation":[105],"without":[106],"model":[107],"retraining,":[108],"(2)":[110],"Token":[112],"Entropy-based":[113],"Early":[114],"Stopping":[115],"strategy":[116],"dynamically":[118],"terminates":[119],"the":[120],"denoising":[121],"process":[122],"based":[123],"on":[124,171],"token-level":[125],"convergence":[126],"metrics.":[127],"We":[128],"implement":[129],"innovations":[131],"through":[132],"specialized":[134],"sparse":[135],"processing":[136],"element":[137],"(PE)":[138],"array":[139],"maximizes":[141],"top-k":[142],"sparsity":[143],"utilization":[144],"minimizing":[146],"idle":[147],"cycles":[148],"row-column":[150],"concatenation,":[151],"complemented":[152],"efficient":[155],"cache":[156],"management":[157],"system":[158],"reduces":[160],"access":[162],"latency":[163],"flexible":[166],"entropybased":[167],"unit.":[169],"Implemented":[170],"U200":[173],"FPGA,":[174],"dLLM-OPU":[175],"achieves":[176],"$2.2":[177],"\\times-5.1":[178],"\\times$":[179,184],"speedup":[180],"$7.6":[182],"\\times-20.3":[183],"energy":[185],"efficiency":[186],"over":[187],"RTX4090":[188],"LLaDA.":[190]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-03-12T00:00:00"}
