{"id":"https://openalex.org/W4401408768","doi":"https://doi.org/10.1145/3673038.3673046","title":"BrickDL: Graph-Level Optimizations for DNNs with Fine-Grained Data Blocking on GPUs","display_name":"BrickDL: Graph-Level Optimizations for DNNs with Fine-Grained Data Blocking on GPUs","publication_year":2024,"publication_date":"2024-08-08","ids":{"openalex":"https://openalex.org/W4401408768","doi":"https://doi.org/10.1145/3673038.3673046"},"language":"en","primary_location":{"id":"doi:10.1145/3673038.3673046","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3673038.3673046","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3673038.3673046","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 53rd International Conference on Parallel Processing","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3673038.3673046","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5041964389","display_name":"Mahesh Lakshminarasimhan","orcid":"https://orcid.org/0000-0002-6813-0536"},"institutions":[{"id":"https://openalex.org/I223532165","display_name":"University of Utah","ror":"https://ror.org/03r0ha626","country_code":"US","type":"education","lineage":["https://openalex.org/I223532165"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mahesh Lakshminarasimhan","raw_affiliation_strings":["University of Utah, United States of America"],"raw_orcid":"https://orcid.org/0000-0002-6813-0536","affiliations":[{"raw_affiliation_string":"University of Utah, United States of America","institution_ids":["https://openalex.org/I223532165"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5030152493","display_name":"Mary Hall","orcid":"https://orcid.org/0000-0002-3058-7573"},"institutions":[{"id":"https://openalex.org/I223532165","display_name":"University of Utah","ror":"https://ror.org/03r0ha626","country_code":"US","type":"education","lineage":["https://openalex.org/I223532165"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mary Hall","raw_affiliation_strings":["University of Utah, United States of America"],"raw_orcid":"https://orcid.org/0000-0002-3058-7573","affiliations":[{"raw_affiliation_string":"University of Utah, United States of America","institution_ids":["https://openalex.org/I223532165"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102746800","display_name":"Samuel Williams","orcid":"https://orcid.org/0000-0002-8327-5717"},"institutions":[{"id":"https://openalex.org/I148283060","display_name":"Lawrence Berkeley National Laboratory","ror":"https://ror.org/02jbv0t02","country_code":"US","type":"facility","lineage":["https://openalex.org/I1330989302","https://openalex.org/I148283060","https://openalex.org/I39565521"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Samuel Williams","raw_affiliation_strings":["Lawrence Berkeley National Laboratory, United States of America"],"raw_orcid":"https://orcid.org/0000-0002-8327-5717","affiliations":[{"raw_affiliation_string":"Lawrence Berkeley National Laboratory, United States of America","institution_ids":["https://openalex.org/I148283060"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5035719658","display_name":"Oscar Antepara","orcid":"https://orcid.org/0000-0002-4596-0289"},"institutions":[{"id":"https://openalex.org/I148283060","display_name":"Lawrence Berkeley National Laboratory","ror":"https://ror.org/02jbv0t02","country_code":"US","type":"facility","lineage":["https://openalex.org/I1330989302","https://openalex.org/I148283060","https://openalex.org/I39565521"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Oscar Antepara","raw_affiliation_strings":["Lawrence Berkeley National Laboratory, United States of America"],"raw_orcid":"https://orcid.org/0000-0002-4596-0289","affiliations":[{"raw_affiliation_string":"Lawrence Berkeley National Laboratory, United States of America","institution_ids":["https://openalex.org/I148283060"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.0668,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.78023634,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"576","last_page":"586"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.9988999962806702,"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"}},"topics":[{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.9988999962806702,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9980000257492065,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.9965000152587891,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7616548538208008},{"id":"https://openalex.org/keywords/blocking","display_name":"Blocking (statistics)","score":0.6898579597473145},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6526877284049988},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.4755147099494934},{"id":"https://openalex.org/keywords/theoretical-computer-science","display_name":"Theoretical computer science","score":0.2510780692100525},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.06659582257270813}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7616548538208008},{"id":"https://openalex.org/C144745244","wikidata":"https://www.wikidata.org/wiki/Q4927286","display_name":"Blocking (statistics)","level":2,"score":0.6898579597473145},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6526877284049988},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.4755147099494934},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2510780692100525},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.06659582257270813}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3673038.3673046","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3673038.3673046","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3673038.3673046","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 53rd International Conference on Parallel Processing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3673038.3673046","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3673038.3673046","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3673038.3673046","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 53rd International Conference on Parallel Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G6859237865","display_name":null,"funder_award_id":"17-SC-20-SC","funder_id":"https://openalex.org/F4320306084","funder_display_name":"U.S. Department of Energy"}],"funders":[{"id":"https://openalex.org/F4320306084","display_name":"U.S. Department of Energy","ror":"https://ror.org/01bj3aw27"}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4401408768.pdf"},"referenced_works_count":42,"referenced_works":["https://openalex.org/W1553808293","https://openalex.org/W1968013322","https://openalex.org/W1976888542","https://openalex.org/W2029833610","https://openalex.org/W2038481157","https://openalex.org/W2039378765","https://openalex.org/W2056971515","https://openalex.org/W2076535184","https://openalex.org/W2103817093","https://openalex.org/W2129471558","https://openalex.org/W2150319905","https://openalex.org/W2160106616","https://openalex.org/W2194775991","https://openalex.org/W2738017032","https://openalex.org/W2804032941","https://openalex.org/W2896180420","https://openalex.org/W2954698171","https://openalex.org/W2962850830","https://openalex.org/W2963616706","https://openalex.org/W2964350391","https://openalex.org/W2972073717","https://openalex.org/W2981758446","https://openalex.org/W2984920043","https://openalex.org/W3012249773","https://openalex.org/W3034107927","https://openalex.org/W3035582633","https://openalex.org/W3120853372","https://openalex.org/W3122286897","https://openalex.org/W3132763358","https://openalex.org/W3152199537","https://openalex.org/W3174529902","https://openalex.org/W3208285274","https://openalex.org/W4212986322","https://openalex.org/W4221003081","https://openalex.org/W4232301837","https://openalex.org/W4241696959","https://openalex.org/W4249932213","https://openalex.org/W4281710230","https://openalex.org/W4318541647","https://openalex.org/W4327930469","https://openalex.org/W6754894021","https://openalex.org/W6950127925"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2748952813","https://openalex.org/W2392835431","https://openalex.org/W2126932387","https://openalex.org/W1965371215","https://openalex.org/W2353762239","https://openalex.org/W2185938410","https://openalex.org/W2484966135","https://openalex.org/W2015050211","https://openalex.org/W2390279801"],"abstract_inverted_index":{"The":[0],"end-to-end":[1],"performance":[2,126,140,165],"of":[3,51,75,78,112,166],"deep":[4,22,39,67,144],"learning":[5,23,40,68,145],"model":[6,127],"inference":[7,69],"is":[8],"often":[9],"limited":[10,45],"by":[11],"excess":[12],"data":[13,19,34,96,101],"movement":[14],"on":[15,103,142,157],"GPUs.":[16,104],"To":[17,58],"reduce":[18],"movement,":[20],"existing":[21,85],"frameworks":[24],"apply":[25],"graph-level":[26,86],"optimizations":[27,43,169],"such":[28],"as":[29,80,147],"operator":[30],"fusion":[31],"to":[32,84,128,149],"exploit":[33],"reuse":[35],"across":[36],"operators":[37],"in":[38],"graphs.":[41],"Such":[42],"are":[44],"and":[46,53,120,123,154,172,178],"cannot":[47],"optimize":[48],"arbitrary":[49],"chains":[50],"compute-":[52],"time-intensive":[54],"operations,":[55],"including":[56],"convolutions.":[57],"address":[59],"these":[60],"limitations,":[61],"this":[62],"paper":[63],"presents":[64],"BrickDL,":[65],"a":[66,76,94,125],"library":[70],"that":[71,98],"implements":[72],"merged":[73,107],"execution":[74,108,136],"sequence":[77],"layers":[79],"an":[81],"orthogonal":[82],"approach":[83],"optimizations.":[87],"BrickDL":[88],"additionally":[89],"employs":[90],"fine-grain":[91],"blocking":[92],"using":[93,114,132],"brick":[95],"layout":[97],"further":[99],"optimizes":[100],"locality":[102],"We":[105,161],"implement":[106],"with":[109,137,170],"the":[110,164,167],"abstraction":[111],"bricks":[113,119,138],"two":[115],"approaches":[116],"\u2013":[117],"padded":[118],"memoized":[121],"bricks,":[122],"develop":[124],"choose":[129],"between":[130],"them":[131],"static":[133],"analysis.":[134],"Merged":[135],"demonstrates":[139],"gains":[141],"well-known":[143],"models":[146],"compared":[148],"PyTorch":[150],"JIT,":[151],"TensorFlow":[152],"XLA,":[153],"cuDNN":[155],"baselines":[156],"NVIDIA":[158],"A100":[159],"GPU.":[160],"also":[162],"characterize":[163],"proposed":[168],"microbenchmarks":[171],"gain":[173],"insights":[174],"into":[175],"their":[176],"applicability":[177],"tradeoffs.":[179]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":1},{"year":2024,"cited_by_count":1}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}