{"id":"https://openalex.org/W4409917015","doi":"https://doi.org/10.1109/pdp66500.2025.00019","title":"Boosting Performance of Iterative Applications on GPUs: Kernel Batching with CUDA Graphs","display_name":"Boosting Performance of Iterative Applications on GPUs: Kernel Batching with CUDA Graphs","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4409917015","doi":"https://doi.org/10.1109/pdp66500.2025.00019"},"language":"en","primary_location":{"id":"doi:10.1109/pdp66500.2025.00019","is_oa":false,"landing_page_url":"https://doi.org/10.1109/pdp66500.2025.00019","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 33rd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP)","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/A5047194408","display_name":"Jonah Ekelund","orcid":null},"institutions":[{"id":"https://openalex.org/I86987016","display_name":"KTH Royal Institute of Technology","ror":"https://ror.org/026vcq606","country_code":"SE","type":"education","lineage":["https://openalex.org/I86987016"]}],"countries":["SE"],"is_corresponding":false,"raw_author_name":"Jonah Ekelund","raw_affiliation_strings":["KTH Royal Institute of Technology,Sweden"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KTH Royal Institute of Technology,Sweden","institution_ids":["https://openalex.org/I86987016"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070572735","display_name":"Stefano Markidis","orcid":null},"institutions":[{"id":"https://openalex.org/I86987016","display_name":"KTH Royal Institute of Technology","ror":"https://ror.org/026vcq606","country_code":"SE","type":"education","lineage":["https://openalex.org/I86987016"]}],"countries":["SE"],"is_corresponding":false,"raw_author_name":"Stefano Markidis","raw_affiliation_strings":["KTH Royal Institute of Technology,Sweden"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KTH Royal Institute of Technology,Sweden","institution_ids":["https://openalex.org/I86987016"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5037069204","display_name":"Ivy Peng","orcid":"https://orcid.org/0000-0003-4158-3583"},"institutions":[{"id":"https://openalex.org/I86987016","display_name":"KTH Royal Institute of Technology","ror":"https://ror.org/026vcq606","country_code":"SE","type":"education","lineage":["https://openalex.org/I86987016"]}],"countries":["SE"],"is_corresponding":false,"raw_author_name":"Ivy Peng","raw_affiliation_strings":["KTH Royal Institute of Technology,Sweden"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KTH Royal Institute of Technology,Sweden","institution_ids":["https://openalex.org/I86987016"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I86987016"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":5,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"70","last_page":"77"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.989300012588501,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.989300012588501,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.977400004863739,"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/T10203","display_name":"Recommender Systems and Techniques","score":0.9761000275611877,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/cuda","display_name":"CUDA","score":0.865093469619751},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7482243776321411},{"id":"https://openalex.org/keywords/kernel","display_name":"Kernel (algebra)","score":0.6911880373954773},{"id":"https://openalex.org/keywords/boosting","display_name":"Boosting (machine learning)","score":0.6901146173477173},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.6832044124603271},{"id":"https://openalex.org/keywords/general-purpose-computing-on-graphics-processing-units","display_name":"General-purpose computing on graphics processing units","score":0.4453742802143097},{"id":"https://openalex.org/keywords/computational-science","display_name":"Computational science","score":0.3447883129119873},{"id":"https://openalex.org/keywords/computer-graphics","display_name":"Computer graphics (images)","score":0.149726003408432},{"id":"https://openalex.org/keywords/graphics","display_name":"Graphics","score":0.14745566248893738},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.13599267601966858},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.08342719078063965}],"concepts":[{"id":"https://openalex.org/C2778119891","wikidata":"https://www.wikidata.org/wiki/Q477690","display_name":"CUDA","level":2,"score":0.865093469619751},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7482243776321411},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.6911880373954773},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.6901146173477173},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.6832044124603271},{"id":"https://openalex.org/C50630238","wikidata":"https://www.wikidata.org/wiki/Q971505","display_name":"General-purpose computing on graphics processing units","level":3,"score":0.4453742802143097},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.3447883129119873},{"id":"https://openalex.org/C121684516","wikidata":"https://www.wikidata.org/wiki/Q7600677","display_name":"Computer graphics (images)","level":1,"score":0.149726003408432},{"id":"https://openalex.org/C21442007","wikidata":"https://www.wikidata.org/wiki/Q1027879","display_name":"Graphics","level":2,"score":0.14745566248893738},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.13599267601966858},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.08342719078063965},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/pdp66500.2025.00019","is_oa":false,"landing_page_url":"https://doi.org/10.1109/pdp66500.2025.00019","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 33rd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Affordable and clean energy","score":0.7699999809265137,"id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":14,"referenced_works":["https://openalex.org/W2080592089","https://openalex.org/W2129960401","https://openalex.org/W3028670792","https://openalex.org/W3092288363","https://openalex.org/W3198396679","https://openalex.org/W4224217616","https://openalex.org/W4321482611","https://openalex.org/W4382202854","https://openalex.org/W4385573061","https://openalex.org/W4385610105","https://openalex.org/W4387030200","https://openalex.org/W4389491867","https://openalex.org/W4403210910","https://openalex.org/W6842583860"],"related_works":["https://openalex.org/W1963859303","https://openalex.org/W2364044215","https://openalex.org/W2389600408","https://openalex.org/W240129890","https://openalex.org/W3048701459","https://openalex.org/W2149078538","https://openalex.org/W2370314112","https://openalex.org/W1912958759","https://openalex.org/W2792081825","https://openalex.org/W2893308117"],"abstract_inverted_index":{"Graphics":[0],"Processing":[1],"Units":[2],"(GPUs)":[3],"have":[4],"become":[5],"the":[6,29,108,145,165,174,181,185,194,206,223],"standard":[7],"in":[8,57,205,218],"accelerating":[9],"scientific":[10],"applications":[11,27,97],"on":[12,189],"heterogeneous":[13],"systems.":[14],"However,":[15],"as":[16,51,55,126],"GPUs":[17],"are":[18],"getting":[19],"faster,":[20],"one":[21,67],"potential":[22],"performance":[23,40,73,104,109,142,182],"bottleneck":[24],"with":[25,180],"GPU-accelerated":[26],"is":[28,160,169],"overhead":[30,112,176],"from":[31,100,113,177,222],"launching":[32],"several":[33,63],"fine-grained":[34],"kernels.":[35,79],"CUDA":[36,94,101,136],"Graph":[37,102],"addresses":[38],"these":[39,90],"challenges":[41],"by":[42,116],"enabling":[43],"a":[44,58,72,93,118,127,141,156,228],"graph-based":[45],"execution":[46],"model":[47],"that":[48,150,193,212],"captures":[49],"operations":[50],"nodes":[52],"and":[53,87,111,138,220,227],"dependence":[54],"edges":[56],"static":[59],"graph.":[60],"Thereby":[61],"consolidating":[62],"kernel":[64,82],"launches":[65,83],"into":[66,84,92],"graph":[68,178,186],"launch.":[69],"We":[70,106],"propose":[71],"optimization":[74],"strategy":[75],"for":[76,103,139,155],"iteratively":[77],"launched":[78],"By":[80],"grouping":[81],"iteration":[85,153,166,196],"batches":[86,91,154],"then":[88],"unrolling":[89,152],"Graph,":[95,137],"iterative":[96,133],"can":[98,215],"benefit":[99],"boosting.":[105],"analyze":[107],"gain":[110,183],"this":[114],"approach":[115],"designing":[117],"skeleton":[119,122,146,207],"application.":[120,208],"The":[121],"application":[123],"also":[124],"serves":[125],"generalized":[128],"example":[129],"of":[130,164,171,184],"converting":[131],"an":[132,161],"solver":[134],"to":[135],"deriving":[140],"model.":[143],"Using":[144],"application,":[147],"we":[148,191,210],"show":[149,192,211],"when":[151],"given":[157],"platform,":[158],"there":[159],"optimal":[162,195],"size":[163,198],"batch,":[167],"which":[168],"independent":[170],"workload,":[172,190],"balancing":[173],"extra":[175],"creation":[179],"execution.":[187],"Depending":[188],"batch":[197],"gives":[199],"more":[200],"than":[201],"$1.4":[202],"\\times$":[203],"speed-up":[204,214],"Furthermore,":[209],"similar":[213],"be":[216],"gained":[217],"Hotspot":[219],"Hotspot3D":[221],"Rodinia":[224],"benchmark":[225],"suite":[226],"Finite-Difference":[229],"Time-Domain":[230],"(FDTD)":[231],"Maxwell":[232],"solver.":[233]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":3}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}