{"id":"https://openalex.org/W4414170448","doi":"https://doi.org/10.1109/iwqos65803.2025.11143255","title":"Deploy Efficient Large Language Model Distributed Inference Pipeline for Heterogeneous GPUs","display_name":"Deploy Efficient Large Language Model Distributed Inference Pipeline for Heterogeneous GPUs","publication_year":2025,"publication_date":"2025-07-02","ids":{"openalex":"https://openalex.org/W4414170448","doi":"https://doi.org/10.1109/iwqos65803.2025.11143255"},"language":"en","primary_location":{"id":"doi:10.1109/iwqos65803.2025.11143255","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iwqos65803.2025.11143255","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/ACM 33rd International Symposium on Quality of Service (IWQoS)","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/A5059592421","display_name":"Junyang Zhang","orcid":"https://orcid.org/0009-0004-8756-4040"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Junyang Zhang","raw_affiliation_strings":["University of Science and Technology of China,Hefei,China"],"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China,Hefei,China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033759584","display_name":"Jiahui Hou","orcid":"https://orcid.org/0000-0002-3340-8585"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiahui Hou","raw_affiliation_strings":["University of Science and Technology of China,Hefei,China"],"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China,Hefei,China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100385166","display_name":"Bowen Zhang","orcid":"https://orcid.org/0000-0003-4640-3086"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bowen Zhang","raw_affiliation_strings":["University of Science and Technology of China,Hefei,China"],"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China,Hefei,China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100341802","display_name":"Xiang\u2010Yang Li","orcid":"https://orcid.org/0000-0002-6070-6625"},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiang-Yang Li","raw_affiliation_strings":["University of Science and Technology of China,Hefei,China"],"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China,Hefei,China","institution_ids":["https://openalex.org/I126520041"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5059592421"],"corresponding_institution_ids":["https://openalex.org/I126520041"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.25453605,"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":"10"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9955999851226807,"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":0.9955999851226807,"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/T10028","display_name":"Topic Modeling","score":0.9955000281333923,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.9916999936103821,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.8111000061035156},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.7961000204086304},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.4977000057697296},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.42750000953674316},{"id":"https://openalex.org/keywords/heuristic","display_name":"Heuristic","score":0.4147999882698059},{"id":"https://openalex.org/keywords/inference-engine","display_name":"Inference engine","score":0.41280001401901245},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.3986999988555908}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8388000130653381},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.8111000061035156},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.7961000204086304},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.4977000057697296},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4465999901294708},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.42750000953674316},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.42289999127388},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.4147999882698059},{"id":"https://openalex.org/C46743427","wikidata":"https://www.wikidata.org/wiki/Q1341685","display_name":"Inference engine","level":3,"score":0.41280001401901245},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.3986999988555908},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.3513999879360199},{"id":"https://openalex.org/C198370458","wikidata":"https://www.wikidata.org/wiki/Q586459","display_name":"Type inference","level":3,"score":0.31139999628067017},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.28690001368522644},{"id":"https://openalex.org/C107027933","wikidata":"https://www.wikidata.org/wiki/Q2006448","display_name":"Stream processing","level":2,"score":0.2745000123977661},{"id":"https://openalex.org/C16320812","wikidata":"https://www.wikidata.org/wiki/Q1812200","display_name":"Idle","level":2,"score":0.272599995136261},{"id":"https://openalex.org/C175309249","wikidata":"https://www.wikidata.org/wiki/Q725864","display_name":"Pipeline transport","level":2,"score":0.26910001039505005},{"id":"https://openalex.org/C129916263","wikidata":"https://www.wikidata.org/wiki/Q1141183","display_name":"Backward chaining","level":4,"score":0.25600001215934753},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.25540000200271606},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.25290000438690186},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2526000142097473}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iwqos65803.2025.11143255","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iwqos65803.2025.11143255","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/ACM 33rd International Symposium on Quality of Service (IWQoS)","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":21,"referenced_works":["https://openalex.org/W2734941459","https://openalex.org/W2969388332","https://openalex.org/W2991040477","https://openalex.org/W3097411828","https://openalex.org/W3130716829","https://openalex.org/W4200071390","https://openalex.org/W4224308101","https://openalex.org/W4285217780","https://openalex.org/W4287391717","https://openalex.org/W4289083201","https://openalex.org/W4289821779","https://openalex.org/W4311887664","https://openalex.org/W4317547647","https://openalex.org/W4321636575","https://openalex.org/W4322718191","https://openalex.org/W4362515116","https://openalex.org/W4376167141","https://openalex.org/W4377371819","https://openalex.org/W4377865168","https://openalex.org/W4386385424","https://openalex.org/W4389521054"],"related_works":[],"abstract_inverted_index":{"The":[0,14,132],"advent":[1],"of":[2,34,57,81,137,155,160],"a":[3,107,112,158],"large":[4],"language":[5],"model":[6],"(LLM)":[7],"has":[8],"revolutionized":[9],"various":[10],"domains":[11],"and":[12,43,74,92,110,150,157],"services.":[13],"inference":[15,36,47,85,101,119,148],"pipeline":[16,59,68,96,120],"system":[17,98,113],"is":[18],"emerging":[19],"as":[20],"an":[21,117,153],"efficient":[22,118],"mechanism":[23],"to":[24,163],"deploy":[25],"LLMs.":[26],"However,":[27],"existing":[28],"works":[29],"barely":[30],"study":[31,91],"the":[32,54,66,82,94,135,145,164],"deployment":[33],"LLM":[35,84,100,147],"on":[37,62,102,121,128],"heterogeneous":[38,103,122,130],"GPUs":[39],"(with":[40],"different":[41,58],"computation":[42],"memory":[44],"capabilities),":[45],"where":[46],"efficiency":[48],"can":[49,78],"be":[50],"heavily":[51],"affected":[52],"by":[53,152],"imbalanced":[55],"performance":[56],"stages.":[60],"Based":[61],"our":[63,138],"empirical":[64],"experience,":[65],"unbalanced":[67],"stages":[69],"incur":[70],"GPU":[71],"wait":[72],"time,":[73],"average":[75,154],"idle":[76],"time":[77],"exceed":[79],"50%":[80],"whole":[83],"process.":[86],"In":[87],"this":[88],"paper,":[89],"we":[90],"optimize":[93],"distributed":[95],"parallelism":[97],"for":[99],"GPUs.":[104,123,131],"We":[105],"present":[106],"heuristic":[108],"algorithm":[109],"implement":[111],"that":[114],"automatically":[115],"deploys":[116],"Extensive":[124],"experiments":[125],"are":[126],"evaluated":[127],"26":[129],"results":[133],"demonstrate":[134],"superiority":[136],"proposed":[139],"system,":[140],"which":[141],"improves":[142],"makespan":[143],"(i.e.,":[144],"total":[146],"latency)":[149],"throughput":[151],"37.1%":[156],"maximum":[159],"83.0%":[161],"compared":[162],"baselines.":[165]},"counts_by_year":[],"updated_date":"2026-03-07T16:01:11.037858","created_date":"2025-10-10T00:00:00"}