{"id":"https://openalex.org/W4212986322","doi":"https://doi.org/10.1145/3503222.3507723","title":"AStitch: enabling a new multi-dimensional optimization space for memory-intensive ML training and inference on modern SIMT architectures","display_name":"AStitch: enabling a new multi-dimensional optimization space for memory-intensive ML training and inference on modern SIMT architectures","publication_year":2022,"publication_date":"2022-02-22","ids":{"openalex":"https://openalex.org/W4212986322","doi":"https://doi.org/10.1145/3503222.3507723"},"language":"en","primary_location":{"id":"doi:10.1145/3503222.3507723","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3503222.3507723","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems","raw_type":"proceedings-article"},"type":"conference-paper","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/A5103042903","display_name":"Zhen Zheng","orcid":"https://orcid.org/0009-0006-2692-713X"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhen Zheng","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005252388","display_name":"Xuanda Yang","orcid":"https://orcid.org/0009-0008-8524-9505"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xuanda Yang","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104030684","display_name":"Pengzhan Zhao","orcid":null},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Pengzhan Zhao","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103437255","display_name":"Guoping Long","orcid":"https://orcid.org/0009-0006-3176-7572"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guoping Long","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088473271","display_name":"Kai Zhu","orcid":"https://orcid.org/0000-0001-5412-5392"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kai Zhu","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112741012","display_name":"Feiwen Zhu","orcid":"https://orcid.org/0009-0005-4813-0685"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Feiwen Zhu","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5052936919","display_name":"Wenyi Zhao","orcid":"https://orcid.org/0000-0001-7308-9542"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wenyi Zhao","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100707681","display_name":"Xiaoyong Liu","orcid":"https://orcid.org/0000-0001-7085-5440"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiaoyong Liu","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100695453","display_name":"Jun Yang","orcid":"https://orcid.org/0000-0002-3365-4989"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jun Yang","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5071200777","display_name":"Jidong Zhai","orcid":"https://orcid.org/0000-0002-7656-6428"},"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":"Jidong Zhai","raw_affiliation_strings":["Tsinghua University, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tsinghua University, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043209884","display_name":"Shuaiwen Leon Song","orcid":"https://orcid.org/0000-0002-8402-1436"},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"The University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Shuaiwen Leon Song","raw_affiliation_strings":["University of Sydney, Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Sydney, Australia","institution_ids":["https://openalex.org/I129604602"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100636012","display_name":"Wei Lin","orcid":"https://orcid.org/0000-0002-3003-0150"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wei Lin","raw_affiliation_strings":["Alibaba Group, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group, China","institution_ids":["https://openalex.org/I45928872"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":6.1023,"has_fulltext":false,"cited_by_count":72,"citation_normalized_percentile":{"value":0.97180617,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"359","last_page":"373"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9991000294685364,"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":0.9991000294685364,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9980999827384949,"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/T10601","display_name":"Handwritten Text Recognition Techniques","score":0.9976999759674072,"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/computer-science","display_name":"Computer science","score":0.8595919609069824},{"id":"https://openalex.org/keywords/compiler","display_name":"Compiler","score":0.661503791809082},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.6175220608711243},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.5768136382102966},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4662562608718872},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4250131845474243},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.36557191610336304},{"id":"https://openalex.org/keywords/theoretical-computer-science","display_name":"Theoretical computer science","score":0.34781765937805176},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.19294533133506775},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.15158501267433167}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8595919609069824},{"id":"https://openalex.org/C169590947","wikidata":"https://www.wikidata.org/wiki/Q47506","display_name":"Compiler","level":2,"score":0.661503791809082},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.6175220608711243},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5768136382102966},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4662562608718872},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4250131845474243},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.36557191610336304},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.34781765937805176},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.19294533133506775},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.15158501267433167}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3503222.3507723","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3503222.3507723","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/9","score":0.44999998807907104,"display_name":"Industry, innovation and infrastructure"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":30,"referenced_works":["https://openalex.org/W2023415862","https://openalex.org/W2106329447","https://openalex.org/W2117689653","https://openalex.org/W2194187530","https://openalex.org/W2412412865","https://openalex.org/W2461316586","https://openalex.org/W2623579764","https://openalex.org/W2765315405","https://openalex.org/W2766166018","https://openalex.org/W2808498336","https://openalex.org/W2905135312","https://openalex.org/W2921898934","https://openalex.org/W2922451262","https://openalex.org/W2961619211","https://openalex.org/W2962745591","https://openalex.org/W2962780374","https://openalex.org/W2979826702","https://openalex.org/W2981758446","https://openalex.org/W3008102594","https://openalex.org/W3087592560","https://openalex.org/W3104493122","https://openalex.org/W3130716829","https://openalex.org/W3148573243","https://openalex.org/W3161395920","https://openalex.org/W3168427063","https://openalex.org/W3174230151","https://openalex.org/W3174529902","https://openalex.org/W3175777295","https://openalex.org/W3175934302","https://openalex.org/W4244917406"],"related_works":["https://openalex.org/W2778498407","https://openalex.org/W2317245370","https://openalex.org/W4249323025","https://openalex.org/W2097707447","https://openalex.org/W198851386","https://openalex.org/W2030310580","https://openalex.org/W947442053","https://openalex.org/W1980160788","https://openalex.org/W2148915962","https://openalex.org/W2283866686"],"abstract_inverted_index":{"This":[0],"work":[1],"reveals":[2],"that":[3,98,163,188],"memory-intensive":[4,106],"computation":[5,122],"is":[6,164],"a":[7,18,93,100,159,215],"rising":[8,88],"performance-critical":[9],"factor":[10],"in":[11,59,229],"recent":[12],"machine":[13,94],"learning":[14,95],"models.":[15],"Due":[16],"to":[17,50,72,166,178,198],"unique":[19],"set":[20],"of":[21,44,62,75,169,194,223],"new":[22,101],"challenges,":[23,89],"existing":[24],"ML":[25,107,151,180,219],"optimizing":[26,96],"compilers":[27],"cannot":[28],"perform":[29],"efficient":[30],"fusion":[31,48,56],"under":[32],"complex":[33,121],"two-level":[34],"dependencies":[35,124],"combined":[36],"with":[37,81,125,221],"just-in-time":[38,143],"demand.":[39],"They":[40],"face":[41],"the":[42,73,201],"dilemma":[43],"either":[45],"performing":[46],"costly":[47],"due":[49,71],"heavy":[51],"redundant":[52],"computation,":[53],"or":[54],"skipping":[55],"which":[57],"results":[58,186],"massive":[60],"number":[61],"kernels.":[63],"Furthermore,":[64],"they":[65],"often":[66],"suffer":[67],"from":[68],"low":[69],"parallelism":[70],"lack":[74],"support":[76,144],"for":[77,105,149,218,231,242],"real-world":[78],"production":[79,208,216],"workloads":[80,220],"irregular":[82],"tensor":[83,134],"shapes.":[84],"To":[85],"address":[86],"these":[87],"we":[90],"propose":[91],"AStitch,":[92],"compiler":[97,161],"opens":[99],"multi-dimensional":[102,117],"optimization":[103,118,183],"space":[104],"computations.":[108],"It":[109],"systematically":[110],"abstracts":[111],"four":[112],"operator-stitching":[113],"schemes":[114],"while":[115],"considering":[116],"objectives,":[119],"tackles":[120],"graph":[123],"novel":[126],"hierarchical":[127],"data":[128],"reuse,":[129],"and":[130,153,182,236],"efficiently":[131],"processes":[132],"various":[133],"shapes":[135],"via":[136],"adaptive":[137],"thread":[138],"mapping.":[139],"Finally,":[140],"AStitch":[141,156,189,213],"provides":[142],"incorporating":[145],"our":[146],"proposed":[147],"optimizations":[148],"both":[150],"training":[152],"inference.":[154],"Although":[155],"serves":[157],"as":[158],"stand-alone":[160],"engine":[162],"portable":[165],"any":[167],"version":[168],"TensorFlow,":[170],"its":[171],"basic":[172],"ideas":[173],"can":[174,190],"be":[175],"generally":[176],"applied":[177],"other":[179],"frameworks":[181],"compilers.":[184],"Experimental":[185],"show":[187],"achieve":[191],"an":[192],"average":[193],"1.84x":[195],"speedup":[196],"(up":[197],"2.73x)":[199],"over":[200],"state-of-the-art":[202],"Google's":[203],"XLA":[204],"solution":[205],"across":[206],"five":[207],"workloads.":[209],"We":[210],"also":[211],"deploy":[212],"onto":[214],"cluster":[217],"thousands":[222],"GPUs.":[224],"The":[225],"system":[226],"has":[227],"been":[228],"operation":[230],"more":[232],"than":[233],"10":[234],"months":[235],"saves":[237],"about":[238],"20,000":[239],"GPU":[240],"hours":[241],"70,000":[243],"tasks":[244],"per":[245],"week.":[246]},"counts_by_year":[{"year":2026,"cited_by_count":6},{"year":2025,"cited_by_count":25},{"year":2024,"cited_by_count":19},{"year":2023,"cited_by_count":17},{"year":2022,"cited_by_count":5}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}