{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,12]],"date-time":"2026-11-12T13:51:29Z","timestamp":1794491489955,"version":"4.9.1"},"reference-count":54,"publisher":"Wiley","issue":"20","license":[{"start":{"date-parts":[[2022,11,8]],"date-time":"2022-11-08T00:00:00Z","timestamp":1667865600000},"content-version":"am","delay-in-days":365,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#am"},{"start":{"date-parts":[[2021,11,8]],"date-time":"2021-11-08T00:00:00Z","timestamp":1636329600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"funder":[{"DOI":"10.13039\/100000015","name":"U.S. Department of Energy","doi-asserted-by":"publisher","award":["DE\u2010AC02\u201006CH11357"],"award-info":[{"award-number":["DE\u2010AC02\u201006CH11357"]}],"id":[{"id":"10.13039\/100000015","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000015","name":"U.S. Department of Energy","doi-asserted-by":"publisher","award":["ECP PROTEAS\u2010TUNE"],"award-info":[{"award-number":["ECP PROTEAS\u2010TUNE"]}],"id":[{"id":"10.13039\/100000015","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Concurrency and Computation"],"published-print":{"date-parts":[[2022,9,10]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:p>We develop a ytopt autotuning framework that leverages Bayesian optimization to explore the parameter space search and compare four different supervised learning methods within Bayesian optimization and evaluate their effectiveness. We select six of the most complex PolyBench benchmarks and apply the newly developed LLVM Clang\/Polly loop optimization pragmas to the benchmarks to optimize them. We then use the autotuning framework to optimize the pragma parameters to improve their performance. The experimental results show that our autotuning approach outperforms the other compiling methods to provide the smallest execution time for the benchmarks syr2k, 3mm, heat\u20103d, lu, and covariance with two large datasets in 200 code evaluations for effectively searching the parameter spaces with up to 170,368 different configurations. We find that the Floyd\u2013Warshall benchmark did not benefit from autotuning. To cope with this issue, we provide some compiler option solutions to improve the performance. Then we present loop autotuning without a user's knowledge using a simple mctree autotuning framework to further improve the performance of the Floyd\u2013Warshall benchmark. We also extend the ytopt autotuning framework to tune a deep learning application.<\/jats:p>","DOI":"10.1002\/cpe.6683","type":"journal-article","created":{"date-parts":[[2021,11,9]],"date-time":"2021-11-09T03:55:02Z","timestamp":1636430102000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":28,"title":["Autotuning PolyBench benchmarks with LLVM Clang\/Polly loop optimization pragmas using Bayesian optimization"],"prefix":"10.1002","volume":"34","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8150-5171","authenticated-orcid":false,"given":"Xingfu","family":"Wu","sequence":"first","affiliation":[{"name":"Mathematics &amp; Computer Science Division Argonne National Laboratory  Lemont Illinois USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Michael","family":"Kruse","sequence":"additional","affiliation":[{"name":"Argonne Leadership Computing Facility Argonne National Laboratory  Lemont Illinois USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Prasanna","family":"Balaprakash","sequence":"additional","affiliation":[{"name":"Mathematics &amp; Computer Science Division Argonne National Laboratory  Lemont Illinois USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hal","family":"Finkel","sequence":"additional","affiliation":[{"name":"Argonne Leadership Computing Facility Argonne National Laboratory  Lemont Illinois USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Paul","family":"Hovland","sequence":"additional","affiliation":[{"name":"Mathematics &amp; Computer Science Division Argonne National Laboratory  Lemont Illinois USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Valerie","family":"Taylor","sequence":"additional","affiliation":[{"name":"Mathematics &amp; Computer Science Division Argonne National Laboratory  Lemont Illinois USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mary","family":"Hall","sequence":"additional","affiliation":[{"name":"Department of Computer Science University of Utah  Salt Lake City Utah USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2021,11,8]]},"reference":[{"key":"e_1_2_11_2_1","doi-asserted-by":"crossref","unstructured":"TapusC ChungI HollingsworthJK.Active harmony: towards automated performance tuning. Proceedings of the 2002 ACM\/IEEE Conference on Supercomputing SC '02; 2002:44.","DOI":"10.1109\/SC.2002.10062"},{"key":"e_1_2_11_3_1","unstructured":"ChungI HollingsworthJK.A case study using automatic performance tuning for large\u2010scale scientific programs. Proceedings of the 15th IEEE International Symposium on High Performance Distributed Computing (HPDC'06); 2006."},{"key":"e_1_2_11_4_1","doi-asserted-by":"publisher","DOI":"10.1002\/cpe.1551"},{"key":"e_1_2_11_5_1","doi-asserted-by":"crossref","unstructured":"TiwariA HollingsworthJK.Online adaptive code generation and tuning. Proceedings of the 2011 IEEE International Parallel & Distributed Processing Symposium (IPDPS'11); 2011.","DOI":"10.1109\/IPDPS.2011.86"},{"key":"e_1_2_11_6_1","doi-asserted-by":"crossref","unstructured":"BalaprakashP GramacyRB WildSM.Active\u2010learning\u2010based surrogate models for empirical performance running. Proceedings of the 2013 IEEE International Conference on Cluster Computing (CLUSTER'13); 2013.","DOI":"10.1109\/CLUSTER.2013.6702683"},{"key":"e_1_2_11_7_1","doi-asserted-by":"publisher","DOI":"10.1002\/cpe.4029"},{"key":"e_1_2_11_8_1","doi-asserted-by":"crossref","unstructured":"OgilvieWF PetoumenosP WangZ LeatherH.Minimizing the cost of iterative compilation with active learning. Proceedings of the 2017 International Symposium on Code Generation and Optimization; 2017.","DOI":"10.1109\/CGO.2017.7863744"},{"key":"e_1_2_11_9_1","doi-asserted-by":"crossref","unstructured":"TiwariA ChenC ChameJ HallM HollingsworthJK.A Scalable auto\u2010tuning framework for compiler optimization. Proceedings of the 23rd IEEE International Parallel and Distributed Computing Symposium (IPDPS'09); 2009; IEEE.","DOI":"10.1109\/IPDPS.2009.5161054"},{"key":"e_1_2_11_10_1","doi-asserted-by":"crossref","unstructured":"MuralidharanS ShantharamM HallM GarlandM CatanzaroB.Nitro: a framework for adaptive code variant tuning. Proceedings of the 2014 IEEE 28th International Parallel and Distributed Processing Symposium (IPDPS'14); 2014.","DOI":"10.1109\/IPDPS.2014.59"},{"key":"e_1_2_11_11_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-28596-8_4"},{"key":"e_1_2_11_12_1","article-title":"Towards autotuning of OpenMP applications on multicore architectures","author":"Katarzynski J","year":"2014","journal-title":"arXiv.org"},{"key":"e_1_2_11_13_1","doi-asserted-by":"crossref","unstructured":"MustafaD AurangzebA EigenmannR.Performance analysis and tuning of automatically parallelized OpenMP applications. Proceedings of the 7th International Conference on OpenMP in the Petascale Era; 2011:151\u2010164; Springer\u2010Verlag Berlin Heidelberg.","DOI":"10.1007\/978-3-642-21487-5_12"},{"key":"e_1_2_11_14_1","unstructured":"ytopt.https:\/\/github.com\/ytopt\u2010team https:\/\/github.com\/ytopt\u2010team\/ytopt https:\/\/github.com\/ytopt\u2010team\/autotune\/blob\/master\/Benchmarks."},{"key":"e_1_2_11_15_1","volume-title":"Machine Learning for HPC Mini\u2010Symposium in PASC19","author":"Balaprakash P","year":"2019"},{"key":"e_1_2_11_16_1","first-page":"2951","article-title":"Practical Bayesian optimization of machine learning algorithms","volume":"2","author":"Snoek J","year":"2012","journal-title":"NIPS'12: Proceedings of the 25th International Conference on Neural Information Processing Systems"},{"key":"e_1_2_11_17_1","doi-asserted-by":"crossref","unstructured":"RoyA BalaprakashP HovlandPD WildSM.Exploiting performance portability in search algorithms for autotuning. Proceedings of the 2016 IEEE International Parallel and Distributed Processing Symposium Workshops: iWAPT; 2016.","DOI":"10.1109\/IPDPSW.2016.85"},{"key":"e_1_2_11_18_1","volume-title":"SC17","author":"Marathe A","year":"2017"},{"key":"e_1_2_11_19_1","doi-asserted-by":"publisher","DOI":"10.1145\/3205289.3205321"},{"key":"e_1_2_11_20_1","doi-asserted-by":"crossref","unstructured":"KruseM FinkelH.User\u2010directed loop\u2010transformations in clang. Proceedings of the 5th Workshop on the LLVM Compiler Infrastructure in HPC (LLVM\u2010HPC 2018. November 12. Dallas); 2018; IEEE.","DOI":"10.1109\/LLVM-HPC.2018.8639402"},{"key":"e_1_2_11_21_1","unstructured":"Clang: A C Language Family Frontend for LLVM.http:\/\/clang.llvm.org"},{"key":"e_1_2_11_22_1","doi-asserted-by":"publisher","DOI":"10.1142\/S0129626412500107"},{"key":"e_1_2_11_23_1","unstructured":"SOLLVE LLVMhttps:\/\/github.com\/SOLLVE\/llvm\u2010project.git"},{"key":"e_1_2_11_24_1","unstructured":"LattnerC.LLVM: An Infrastructure for Multi\u2010Stage Optimization. Master's thesis. Computer Science Department University of Illinois: 2002."},{"key":"e_1_2_11_25_1","doi-asserted-by":"crossref","unstructured":"WuX KruseM BalaprakashP et al.Autotuning PolyBench benchmarks with LLVM clang\/polly loop optimization pragmas using Bayesian optimization. Proceedings of SC20 Workshop on Performance Modeling Benchmarking and Simulation of High Performance Computer Systems; 2020; IEEE Computer Society Washington DC.","DOI":"10.1109\/PMBS51919.2020.00012"},{"key":"e_1_2_11_26_1","unstructured":"YukiT PouchetL.PolyBench 4.2; May 9 2016.http:\/\/sourceforge.net\/projects\/polybench\/"},{"key":"e_1_2_11_27_1","unstructured":"HeadT MechCoderGL ShcherbatyiI.scikit\u2010optimize\/scikit\u2010optimize: v0.5.2; 2018."},{"key":"e_1_2_11_28_1","unstructured":"LindauerM EggenspergerK FeurerM et al.BOAH: a tool suite for multi\u2010fidelity bayesian optimization & analysis of hyperparameters; 2019. arXiv:1908.06756 [cs.LG]."},{"key":"e_1_2_11_29_1","unstructured":"autoML.https:\/\/www.automl.org"},{"key":"e_1_2_11_30_1","doi-asserted-by":"publisher","DOI":"10.1145\/2925987"},{"key":"e_1_2_11_31_1","unstructured":"BondhugulaU.High performance code generation in MLIR: an early case study with GEMM; March 2020. arXiv:2003.00532v1 [cs.PF]."},{"key":"e_1_2_11_32_1","unstructured":"mctree.https:\/\/github.com\/Meinersbur\/mctree."},{"key":"e_1_2_11_33_1","doi-asserted-by":"crossref","unstructured":"KruseM FinkelH WuX.Autotuning search space for loop transformations. Proceedings of SC20 Workshop on the LLVM Compiler Infrastructure in HPC; 2020; IEEE Computer Society Washington DC.","DOI":"10.1109\/LLVMHPCHiPar51896.2020.00007"},{"key":"e_1_2_11_34_1","unstructured":"Keras.https:\/\/keras.io\/about\/."},{"key":"e_1_2_11_35_1","unstructured":"TensorFlow.https:\/\/www.tensorflow.org."},{"key":"e_1_2_11_36_1","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2018.2841200"},{"key":"e_1_2_11_37_1","unstructured":"WhaleyRC DongarraJJ.Automatically tuned linear algebra software. Proceedings of the 1998 ACM\/IEEE Conference on Supercomputing; 1998:1\u201027; IEEE Computer Society Washington DC."},{"key":"e_1_2_11_38_1","unstructured":"ChenC ChameJ HallM.Combining models and guided empirical search to optimize for multiple levels of the memory hierarchy. Proceedings of the International Symposium on Code Generation and Optimization; 2005."},{"key":"e_1_2_11_39_1","doi-asserted-by":"publisher","DOI":"10.1145\/3197978"},{"key":"e_1_2_11_40_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-13374-9_21"},{"key":"e_1_2_11_41_1","doi-asserted-by":"crossref","unstructured":"BruelP Quinito MasnadaS VideauB LegrandA VincentJM GoldmanA.Autotuning under tight budget constraints: a transparent design of experiments approach. Proceedings of the International Symposium in Cluster Cloud and Grid Computing (CCGrid'19); 2019.","DOI":"10.1109\/CCGRID.2019.00026"},{"key":"e_1_2_11_42_1","doi-asserted-by":"crossref","unstructured":"HartonoA NorrisB SadayappanP.Annotation\u2010based empirical performance tuning using Orio. Proceedings of the 2009 IEEE International Symposium on Parallel Distributed Processing; 2009.","DOI":"10.1109\/IPDPS.2009.5161004"},{"key":"e_1_2_11_43_1","doi-asserted-by":"crossref","unstructured":"AnselJ KamilS VeeramachaneniK et al.OpenTuner: an extensible framework for program autotuning. Proceedings of the 23rd International Conference on Parallel Architectures and Compilation Techniques; 2014:303\u2010316; ACM.","DOI":"10.1145\/2628071.2628092"},{"key":"e_1_2_11_44_1","doi-asserted-by":"crossref","unstructured":"NelsonT RiveraA BalaprakashP et al.Generating efficient tensor contractions for GPUs. Proceedings of the 2015 44th International Conference on Parallel Processing; 2015.","DOI":"10.1109\/ICPP.2015.106"},{"key":"e_1_2_11_45_1","unstructured":"ZingalesGPET.HalideTuner: Generating and Tuning Halide Schedules with OpenTuner. Master's thesis. Massachusetts Institute of Technology; 2015."},{"key":"e_1_2_11_46_1","doi-asserted-by":"crossref","unstructured":"RaschA HaidlM GorlatchS.ATF: a generic auto\u2010tuning framework. Proceedings of the 19th International Conference on High Performance Computing and Communications; 15th International Conference on Smart City; 3rd International Conference on Data Science and Systems (HPCC\/SmartCity\/DSS); 2017:64\u201071; IEEE.","DOI":"10.1109\/HPCC-SmartCity-DSS.2017.9"},{"key":"e_1_2_11_47_1","volume-title":"arXiv:1703.06503","author":"Nugteren C","year":"2017"},{"key":"e_1_2_11_48_1","doi-asserted-by":"crossref","unstructured":"PfaffeP GrosserT TillmannM.Efficient hierarchical online\u2010autotuning: a case study on polyhedral accelerator mapping. Proceedings of the ACM International Conference on Supercomputing; 2019; Association for Computing Machinery New York NY.","DOI":"10.1145\/3330345.3330377"},{"key":"e_1_2_11_49_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2018.08.004"},{"key":"e_1_2_11_50_1","unstructured":"Haj\u2010AliA GencH HuangQ et al.ProTuner: tuning programs with Monte Carlo tree search; 2020. arXiv preprint."},{"key":"e_1_2_11_51_1","doi-asserted-by":"publisher","DOI":"10.1145\/3427093"},{"key":"e_1_2_11_52_1","doi-asserted-by":"publisher","DOI":"10.1145\/3306346.3322967"},{"key":"e_1_2_11_53_1","unstructured":"BeaugnonU ClementB TollenaereN CohenA.On the representation of partially specified implementations and its application to the optimization of linear algebra kernels on GPU; 2019. arXiv preprint; 2019. arXiv:1904.03383 [cs.PL]."},{"key":"e_1_2_11_54_1","doi-asserted-by":"crossref","unstructured":"SilvanoC PalermoG AgostaG et al.Autotuning and adaptivity in energy efficient HPC systems: the ANTAREX toolbox. Proceedings of the 15th ACM International Conference on Computing Frontiers; 2018:270\u2010275; ACM New York NY.","DOI":"10.1145\/3203217.3205338"},{"key":"e_1_2_11_55_1","doi-asserted-by":"crossref","unstructured":"WuX MaratheA JanaS et al.Toward an end\u2010to\u2010end auto\u2010tuning framework in HPC PowerStack. Proceedings of Energy Efficient HPC State of Practice 2020 (EE HPC SOP 20); 2020; IEEE Computer Society Washington DC.","DOI":"10.1109\/CLUSTER49012.2020.00068"}],"container-title":["Concurrency and Computation: Practice and Experience"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/cpe.6683","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/cpe.6683","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/am-pdf\/10.1002\/cpe.6683","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/cpe.6683","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,4,10]],"date-time":"2025-04-10T00:24:26Z","timestamp":1744244666000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1002\/cpe.6683"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,11,8]]},"references-count":54,"journal-issue":{"issue":"20","published-print":{"date-parts":[[2022,9,10]]}},"alternative-id":["10.1002\/cpe.6683"],"URL":"https:\/\/doi.org\/10.1002\/cpe.6683","archive":["Portico"],"relation":{},"ISSN":["1532-0626","1532-0634"],"issn-type":[{"value":"1532-0626","type":"print"},{"value":"1532-0634","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,11,8]]},"assertion":[{"value":"2021-02-26","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-10-06","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-11-08","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e6683"}}