{"id":"https://openalex.org/W7160899795","doi":"https://doi.org/10.48550/arxiv.2605.08835","title":"SynerDiff: Synergetic Continuous Batching for Fast and Parallel Diffusion Model Inference","display_name":"SynerDiff: Synergetic Continuous Batching for Fast and Parallel Diffusion Model Inference","publication_year":2026,"publication_date":"2026-05-09","ids":{"openalex":"https://openalex.org/W7160899795","doi":"https://doi.org/10.48550/arxiv.2605.08835"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.08835","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08835","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.08835","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5089387872","display_name":"Ziqi Zhou","orcid":"https://orcid.org/0000-0003-0427-7819"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Ziqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135954324","display_name":"Peng Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Peng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135941545","display_name":"Yuxin Liang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liang, Yuxin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054192954","display_name":"Mingliu Liu","orcid":"https://orcid.org/0000-0002-4101-3938"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Mingliu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5102003456","display_name":"Jia Lu","orcid":"https://orcid.org/0000-0001-5084-818X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Jia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11165","display_name":"Image and Video Quality Assessment","score":0.2743000090122223,"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/T11165","display_name":"Image and Video Quality Assessment","score":0.2743000090122223,"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/T11478","display_name":"Caching and Content Delivery","score":0.1264999955892563,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T13553","display_name":"Age of Information Optimization","score":0.08940000087022781,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/scheduling","display_name":"Scheduling (production processes)","score":0.6197999715805054},{"id":"https://openalex.org/keywords/queue","display_name":"Queue","score":0.6072999835014343},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.517799973487854},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.37779998779296875},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.326200008392334},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.28450000286102295}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8302000164985657},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6197999715805054},{"id":"https://openalex.org/C160403385","wikidata":"https://www.wikidata.org/wiki/Q220543","display_name":"Queue","level":2,"score":0.6072999835014343},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5328999757766724},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.517799973487854},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.4514000117778778},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.37779998779296875},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.326200008392334},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.28450000286102295},{"id":"https://openalex.org/C107568181","wikidata":"https://www.wikidata.org/wiki/Q5319000","display_name":"Dynamic priority scheduling","level":3,"score":0.2818000018596649},{"id":"https://openalex.org/C181889124","wikidata":"https://www.wikidata.org/wiki/Q380204","display_name":"Service level","level":2,"score":0.2554999887943268},{"id":"https://openalex.org/C19012869","wikidata":"https://www.wikidata.org/wiki/Q578372","display_name":"Response time","level":2,"score":0.251800000667572}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.08835","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08835","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.08835","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08835","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/8","display_name":"Decent work and economic growth","score":0.4231911599636078}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"expansion":[1],"of":[2],"Artificial":[3],"Intelligence-generated":[4],"content":[5],"service":[6],"requires":[7],"diffusion":[8],"model":[9],"serving":[10],"to":[11,36,99,112,134,159,162],"simultaneously":[12],"achieve":[13],"high":[14,166],"throughput":[15,48,145],"and":[16,49,88,108,148,153],"low":[17],"task":[18],"end-to-end":[19],"(E2E)":[20],"latency.":[21],"However,":[22],"existing":[23],"continuous":[24,64],"batching":[25,65],"methods":[26],"suffer":[27],"from":[28],"severe":[29],"resource":[30,78,83],"contention":[31,79],"during":[32],"UNet-VAE":[33],"concurrency,":[34],"leading":[35],"latency":[37,51,115],"spikes.":[38],"Furthermore,":[39],"concurrent":[40,106],"multi-task":[41],"scheduling":[42,54,100],"entails":[43],"a":[44,102,123],"trade-off":[45],"between":[46],"UNet":[47,118],"VAE":[50,86,114],"across":[52],"varying":[53],"strategies.":[55],"To":[56],"address":[57],"these,":[58],"we":[59],"propose":[60],"SynerDiff,":[61],"an":[62],"efficient":[63],"system":[66,136],"built":[67],"on":[68,131],"intra-inter":[69],"level":[70],"synergy.":[71],"At":[72,91],"the":[73,92],"intra-concurrency":[74,110],"level,":[75,94],"SynerDiff":[76,143],"alleviates":[77],"by":[80,146,157],"pruning":[81],"component-specific":[82],"bottlenecks":[84],"via":[85],"Chunking":[87],"Adaptive":[89],"Skip-CFG.":[90],"inter-concurrency":[93],"leveraging":[95],"components'":[96],"differential":[97],"sensitivity":[98],"granularities,":[101],"threshold-aware":[103],"scheduler":[104],"plans":[105],"sequences":[107],"tunes":[109],"decisions":[111],"minimize":[113],"while":[116,164],"maintaining":[117],"within":[119],"high-throughput":[120],"threshold.":[121],"Additionally,":[122],"feedback":[124],"controller":[125],"dynamically":[126],"adjusts":[127],"this":[128],"threshold":[129],"based":[130],"queue":[132],"loads":[133],"boost":[135],"capacity":[137],"ceiling.":[138],"Experimental":[139],"results":[140],"show":[141],"that,":[142],"improves":[144],"1.6$\\times$":[147],"decreases":[149],"both":[150],"average":[151],"E2E":[152],"P99":[154],"tail":[155],"latencies":[156],"up":[158],"78.7\\%,":[160],"compared":[161],"benchmarks":[163],"guaranteeing":[165],"image":[167],"fidelity.":[168]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-13T00:00:00"}