{"id":"https://openalex.org/W4417243891","doi":"https://doi.org/10.1145/3760250.3762228","title":"XY-Serve: End-to-End Versatile Production Serving for Dynamic LLM Workloads","display_name":"XY-Serve: End-to-End Versatile Production Serving for Dynamic LLM Workloads","publication_year":2025,"publication_date":"2025-12-11","ids":{"openalex":"https://openalex.org/W4417243891","doi":"https://doi.org/10.1145/3760250.3762228"},"language":null,"primary_location":{"id":"doi:10.1145/3760250.3762228","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3760250.3762228","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 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3760250.3762228","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5000111354","display_name":"Mingcong Song","orcid":"https://orcid.org/0009-0002-5289-685X"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Mingcong Song","raw_affiliation_strings":["Huawei Technologies Co., Ltd., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0002-5289-685X","affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd., Beijing, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103213494","display_name":"Xinru Tang","orcid":"https://orcid.org/0009-0004-6038-3709"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xinru Tang","raw_affiliation_strings":["Tsinghua University, BNRist, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0004-6038-3709","affiliations":[{"raw_affiliation_string":"Tsinghua University, BNRist, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026136192","display_name":"Feng Hou","orcid":"https://orcid.org/0009-0006-6937-9057"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fengfan Hou","raw_affiliation_strings":["Huawei Technologies Co., Ltd., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0006-6937-9057","affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd., Beijing, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092305334","display_name":"Jing Li","orcid":"https://orcid.org/0009-0005-3806-6499"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jing Li","raw_affiliation_strings":["Huawei Technologies Co., Ltd., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0005-3806-6499","affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd., Beijing, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Wei Wei","orcid":"https://orcid.org/0009-0003-1991-2253"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wei Wei","raw_affiliation_strings":["Huawei Technologies Co., Ltd., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0003-1991-2253","affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd., Beijing, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109785533","display_name":"Yipeng Ma","orcid":"https://orcid.org/0009-0002-3826-2180"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yipeng Ma","raw_affiliation_strings":["Huawei Technologies Co., Ltd., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0002-3826-2180","affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd., Beijing, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009609913","display_name":"Runqiu Xiao","orcid":"https://orcid.org/0000-0003-1121-3709"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Runqiu Xiao","raw_affiliation_strings":["Huawei Technologies Co., Ltd., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0005-3550-5636","affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd., Beijing, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101107776","display_name":"Huamin Si","orcid":"https://orcid.org/0009-0006-7511-6624"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hongjie Si","raw_affiliation_strings":["Huawei Technologies Co., Ltd., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0006-7511-6624","affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd., Beijing, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102699169","display_name":"Dingcheng Jiang","orcid":"https://orcid.org/0009-0004-4379-694X"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dingcheng Jiang","raw_affiliation_strings":["Tsinghua University, BNRist, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0004-4379-694X","affiliations":[{"raw_affiliation_string":"Tsinghua University, BNRist, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054524841","display_name":"Shouyi Yin","orcid":"https://orcid.org/0000-0003-2309-572X"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shouyi Yin","raw_affiliation_strings":["Tsinghua University, BNRist, Beijing, China and Shanghai AI Laboratory, Shanghai, China"],"raw_orcid":"https://orcid.org/0000-0003-2309-572X","affiliations":[{"raw_affiliation_string":"Tsinghua University, BNRist, Beijing, China and Shanghai AI Laboratory, Shanghai, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066434078","display_name":"Yang Hu","orcid":"https://orcid.org/0000-0001-6942-4395"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang Hu","raw_affiliation_strings":["Tsinghua University, BNRist, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0001-6942-4395","affiliations":[{"raw_affiliation_string":"Tsinghua University, BNRist, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5103437255","display_name":"Guoping Long","orcid":"https://orcid.org/0009-0006-3176-7572"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guoping Long","raw_affiliation_strings":["Huawei Technologies Co., Ltd., Beijing, China"],"raw_orcid":"https://orcid.org/0009-0006-3176-7572","affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd., Beijing, China","institution_ids":["https://openalex.org/I2250955327"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":12,"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":0.39105236,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"314","last_page":"329"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.42160001397132874,"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":0.42160001397132874,"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":0.2554999887943268,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.0357000008225441,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/abstraction","display_name":"Abstraction","score":0.7039999961853027},{"id":"https://openalex.org/keywords/workload","display_name":"Workload","score":0.5695000290870667},{"id":"https://openalex.org/keywords/generality","display_name":"Generality","score":0.5440999865531921},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.5188000202178955},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.48489999771118164},{"id":"https://openalex.org/keywords/padding","display_name":"Padding","score":0.42089998722076416},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.4056999981403351},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.39649999141693115}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8985999822616577},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.7039999961853027},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.571399986743927},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.5695000290870667},{"id":"https://openalex.org/C2780767217","wikidata":"https://www.wikidata.org/wiki/Q5532421","display_name":"Generality","level":2,"score":0.5440999865531921},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.5188000202178955},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.48489999771118164},{"id":"https://openalex.org/C165435473","wikidata":"https://www.wikidata.org/wiki/Q1509884","display_name":"Padding","level":2,"score":0.42089998722076416},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.4056999981403351},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.39649999141693115},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3765999972820282},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.3637999892234802},{"id":"https://openalex.org/C41045048","wikidata":"https://www.wikidata.org/wiki/Q202843","display_name":"Linear programming","level":2,"score":0.36000001430511475},{"id":"https://openalex.org/C67953723","wikidata":"https://www.wikidata.org/wiki/Q192525","display_name":"Workstation","level":2,"score":0.33379998803138733},{"id":"https://openalex.org/C2778348673","wikidata":"https://www.wikidata.org/wiki/Q739302","display_name":"Production (economics)","level":2,"score":0.3116999864578247},{"id":"https://openalex.org/C42812","wikidata":"https://www.wikidata.org/wiki/Q1082910","display_name":"Partition (number theory)","level":2,"score":0.304500013589859},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.2825999855995178},{"id":"https://openalex.org/C37404715","wikidata":"https://www.wikidata.org/wiki/Q380679","display_name":"Dynamic programming","level":2,"score":0.273499995470047},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.27219998836517334},{"id":"https://openalex.org/C93996380","wikidata":"https://www.wikidata.org/wiki/Q44127","display_name":"Server","level":2,"score":0.2712000012397766},{"id":"https://openalex.org/C34165917","wikidata":"https://www.wikidata.org/wiki/Q188267","display_name":"Programming paradigm","level":2,"score":0.26409998536109924},{"id":"https://openalex.org/C88196245","wikidata":"https://www.wikidata.org/wiki/Q8034984","display_name":"Working set","level":2,"score":0.2581000030040741},{"id":"https://openalex.org/C25344961","wikidata":"https://www.wikidata.org/wiki/Q192726","display_name":"Virtual machine","level":2,"score":0.25369998812675476},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.2508000135421753}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3760250.3762228","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3760250.3762228","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 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3760250.3762228","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3760250.3762228","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 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":9,"referenced_works":["https://openalex.org/W2160428323","https://openalex.org/W2954698171","https://openalex.org/W2982219368","https://openalex.org/W3096609285","https://openalex.org/W3157657667","https://openalex.org/W4327911434","https://openalex.org/W4387321091","https://openalex.org/W4395112660","https://openalex.org/W4402671542"],"related_works":[],"abstract_inverted_index":{"Meeting":[0],"growing":[1],"demands":[2],"for":[3,109,181],"low":[4],"latency":[5],"and":[6,24,186],"cost":[7],"efficiency":[8,46],"in":[9],"production-grade":[10],"large":[11],"language":[12],"model":[13],"(LLM)":[14],"serving":[15],"systems":[16],"requires":[17],"integrating":[18],"advanced":[19],"optimization":[20],"techniques.":[21],"However,":[22],"dynamic":[23,137],"unpredictable":[25],"input-output":[26],"lengths":[27],"of":[28,37,103,120,201],"LLM,":[29],"compounded":[30],"by":[31,85],"these":[32],"optimizations,":[33],"exacerbate":[34],"the":[35,82,101,117,199],"issues":[36],"workload":[38,83],"variability,":[39],"making":[40],"it":[41],"difficult":[42],"to":[43,136,160,193],"maintain":[44],"high":[45],"on":[47,171,204],"AI":[48],"accelerators,":[49],"especially":[50],"DSAs":[51],"with":[52,122,146,154,166],"tile-based":[53],"programming":[54],"models.":[55],"To":[56],"address":[57],"this":[58,106],"challenge,":[59],"we":[60,111,128,197],"introduce":[61,129],"XY-Serve,":[62],"a":[63,113,130,177],"versatile,":[64],"Ascend":[65,172],"NPU":[66],"native,":[67],"end-to-end":[68,162],"production":[69],"LLM-serving":[70],"system.":[71],"The":[72],"core":[73],"idea":[74],"is":[75],"an":[76],"abstraction":[77,107],"mechanism":[78],"that":[79,115,134],"smooths":[80],"out":[81],"variability":[84],"decomposing":[86],"computations":[87],"into":[88],"unified,":[89],"hardware-friendly,":[90],"fine-grained":[91],"meta":[92],"primitives.":[93],"Then,":[94],"kernels":[95,191],"can":[96],"efficiently":[97],"execute":[98],"without":[99],"concerning":[100],"irregularity":[102],"workload.":[104],"After":[105],"mechanism,":[108],"Attention,":[110],"propose":[112],"meta-kernel":[114],"computes":[116],"basic":[118],"pattern":[119],"GEMM-Softmax-GEMM":[121],"architectural-aware":[123],"tile":[124,149],"sizes.":[125,150],"For":[126],"Linear,":[127],"virtual":[131],"padding":[132],"scheme":[133],"adapts":[135],"shape":[138],"changes":[139],"while":[140],"using":[141],"highly":[142],"efficient":[143],"GEMM":[144],"primitives":[145],"assorted":[147],"fixed":[148],"XY-Serve":[151],"sits":[152],"harmoniously":[153],"vLLM.":[155],"Experimental":[156],"results":[157],"show":[158],"up":[159],"95%":[161],"throughput":[163],"improvement":[164],"compared":[165],"current":[167],"publicly":[168],"available":[169],"baselines":[170],"NPUs.":[173],"We":[174],"also":[175],"set":[176],"new":[178],"performance":[179],"record":[180],"Linear":[182],"(average":[183,188],"14.6%":[184],"faster)":[185,190],"Attention":[187],"21.5%":[189],"relative":[192],"existing":[194],"libraries.":[195],"Lastly,":[196],"demonstrate":[198],"generality":[200],"our":[202],"technologies":[203],"GPU":[205],"platform.":[206]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-12-11T00:00:00"}