{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,2]],"date-time":"2025-11-02T17:16:18Z","timestamp":1762103778930,"version":"build-2065373602"},"reference-count":37,"publisher":"Wiley","issue":"16","license":[{"start":{"date-parts":[[2024,4,15]],"date-time":"2024-04-15T00:00:00Z","timestamp":1713139200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"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"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Concurrency and Computation"],"published-print":{"date-parts":[[2024,7,25]]},"abstract":"<jats:title>Summary<\/jats:title><jats:p>Parallel I\/O is a critical technique for moving data between compute and storage subsystems of supercomputers. With massive amounts of data produced or consumed by compute nodes, high\u2010performant parallel I\/O is essential. I\/O benchmarks play an important role in this process; however, there is a scarcity of I\/O benchmarks representative of current workloads on HPC systems. Toward creating representative I\/O kernels from real\u2010world applications, we have created <jats:styled-content>h5bench<\/jats:styled-content>\u00a0, a set of I\/O kernels that exercise hierarchical data format version 5 (HDF5) I\/O on parallel file systems in numerous dimensions. Our focus on HDF5 is due to the parallel I\/O library's heavy usage in various scientific applications running on supercomputing systems. The various tests benchmarked in the <jats:styled-content>h5bench<\/jats:styled-content>\u00a0suite include I\/O operations (read and write), data locality (arrays of basic data types and arrays of structures), array dimensionality (one\u2010dimensional arrays, two\u2010dimensional meshes, three\u2010dimensional cubes), I\/O modes (synchronous and asynchronous). In this paper, we present the observed performance of <jats:styled-content>h5bench<\/jats:styled-content>\u00a0executed along several of these dimensions on existing supercomputers (Cori and Summit) and pre\u2010exascale platforms (Perlmutter, Theta, and Polaris). <jats:styled-content>h5bench<\/jats:styled-content>\u00a0measurements can be used to identify performance bottlenecks and their root causes and evaluate I\/O optimizations. As the I\/O patterns of <jats:styled-content>h5bench<\/jats:styled-content>\u00a0are diverse and capture the I\/O behaviors of various HPC applications, this study will be helpful to the broader supercomputing and I\/O community.<\/jats:p>","DOI":"10.1002\/cpe.8046","type":"journal-article","created":{"date-parts":[[2024,4,16]],"date-time":"2024-04-16T04:30:09Z","timestamp":1713241809000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["h5bench: A unified benchmark suite for evaluating HDF5 I\/O performance on pre\u2010exascale platforms"],"prefix":"10.1002","volume":"36","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-3915-1135","authenticated-orcid":false,"given":"Jean Luca","family":"Bez","sequence":"first","affiliation":[{"name":"Scientific Data Division Lawrence Berkeley National Laboratory  Berkeley California USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Houjun","family":"Tang","sequence":"additional","affiliation":[{"name":"Scientific Data Division Lawrence Berkeley National Laboratory  Berkeley California USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Scot","family":"Breitenfeld","sequence":"additional","affiliation":[{"name":"HDF Group  Champaign Illinois USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Huihuo","family":"Zheng","sequence":"additional","affiliation":[{"name":"Data Science Group Argonne National Laboratory  Lemont Illinois USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Wei\u2010Keng","family":"Liao","sequence":"additional","affiliation":[{"name":"Electrical and Computer Engineering Department Northwestern University  Evanston Illinois USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kaiyuan","family":"Hou","sequence":"additional","affiliation":[{"name":"Electrical and Computer Engineering Department Northwestern University  Evanston Illinois USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zanhua","family":"Huang","sequence":"additional","affiliation":[{"name":"Computer Science Department Northwestern University  Evanston Illinois United States"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Suren","family":"Byna","sequence":"additional","affiliation":[{"name":"Scientific Data Division Lawrence Berkeley National Laboratory  Berkeley California USA"},{"name":"Department of Computer Science and Engineering The Ohio State University  Ohio United States"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","published-online":{"date-parts":[[2024,4,15]]},"reference":[{"key":"e_1_2_9_2_1","unstructured":"The HDF Group.Hierarchical data format. version 51997.http:\/\/www.hdfgroup.org\/HDF5"},{"key":"e_1_2_9_3_1","doi-asserted-by":"crossref","unstructured":"FolkM HeberG KoziolQ PourmalE RobinsonD.An overview of the HDF5 technology suite and its applications. Paper presented at: EDBT\/ICDT. Association for Computing Machinery. Uppsala Sweden;2011:36\u201047. doi:10.1145\/1966895.1966900","DOI":"10.1145\/1966895.1966900"},{"key":"e_1_2_9_4_1","doi-asserted-by":"crossref","first-page":"76","DOI":"10.1109\/38.56302","article-title":"netCDF: an interface for scientific data access","volume":"10","author":"Rew R","journal-title":"IEEE Comput Graph Appl"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cpc.2009.08.005"},{"key":"e_1_2_9_6_1","doi-asserted-by":"crossref","unstructured":"LiJ LiaokW ChoudharyA et al.Parallel netCDF: a high\u2010performance scientific I\/O Interface. Paper presented at: SC'03: Proceedings of the 2003 ACM\/IEEE Conference on Supercomputing. IEEE. Phoenix AZ USA;2003:39\u201039. doi:10.1145\/1048935.1050189","DOI":"10.1145\/1048935.1050189"},{"key":"e_1_2_9_7_1","first-page":"23","volume-title":"Icpads","author":"Thakur R","year":"1999"},{"key":"e_1_2_9_8_1","unstructured":"KunkelJM BentJ KunkelJ KunkelGS.Establishing the IO\u2010500 Benchmark.2016.https:\/\/www.vi4io.org\/_media\/io500\/about\/io500\u2010establishing.pdf"},{"key":"e_1_2_9_9_1","doi-asserted-by":"crossref","unstructured":"TangH KoziolQ BynaS MainzerJ LiT.Enabling transparent asynchronous I\/O using background threads. Paper presented at: 2019 IEEE\/ACM Fourth International Parallel Data Systems Workshop (PDSW). IEEE. Denver CO USA;2019:11\u201019. doi:10.1109\/PDSW49588.2019.00006","DOI":"10.1109\/PDSW49588.2019.00006"},{"key":"e_1_2_9_10_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2021.3090322"},{"key":"e_1_2_9_11_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11390-020-9822-9"},{"key":"e_1_2_9_12_1","unstructured":"MillerMC.Multi\u2010purpose application\u2010centric scalable I\/O proxy application. Version 002015.https:\/\/www.osti.gov\/biblio\/1232293"},{"key":"e_1_2_9_13_1","unstructured":"BynaS HowisonM.Parallel I\/O kernel (PIOK) suite.2015.https:\/\/sdm.lbl.gov\/exahdf5\/software.html"},{"key":"e_1_2_9_14_1","unstructured":"Computational Science fFC.The FLASH code.http:\/\/flash.uchicago.edu\/site\/flashcode\/"},{"key":"e_1_2_9_15_1","unstructured":"ColellaP GravesDT JohnsonJN et al.Chombo software package for AMR applications design document. Technical Report No. LBNL\u20106616E. Lawrence Berkeley National Laboratory2003."},{"key":"e_1_2_9_16_1","doi-asserted-by":"publisher","DOI":"10.21105\/joss.01370"},{"key":"e_1_2_9_17_1","doi-asserted-by":"crossref","unstructured":"ZhengH VishwanathV KoziolQ et al.HDF5 cache VOL: efficient and scalable parallel I\/O through caching data on node\u2010local storage. Paper presented at: 2022 22nd IEEE International Symposium on Cluster Cloud and Internet Computing (CCGrid). IEEE. Taormina Italy;2022:61\u201070. doi:10.1109\/CCGrid54584.2022.00015","DOI":"10.1109\/CCGrid54584.2022.00015"},{"volume-title":"Proceedings of Cray User Group Meeting, CUG 2016","year":"2016","author":"Henseler D","key":"e_1_2_9_18_1"},{"key":"e_1_2_9_19_1","unstructured":"HDF5.Virtual object layer (VOL) user guide.2020.https:\/\/github.com\/HDFGroup\/hdf5doc\/blob\/master\/RFCs\/HDF5\/VOL\/connector_author_guide\/vol_connector_author_guide.pdf"},{"key":"e_1_2_9_20_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2017.2766062"},{"key":"e_1_2_9_21_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-540-75416-9_34"},{"key":"e_1_2_9_22_1","unstructured":"HerouxMA McInnesLC ThakurR et al.ECP Software Technology Capability Assessment Report. Technical Report. USDOE Office of Science (SC)2020."},{"key":"e_1_2_9_23_1","unstructured":"WuK BynaS DongB.USDOE. VPIC IO utilities.2018.https:\/\/www.osti.gov\/biblio\/1487266"},{"key":"e_1_2_9_24_1","unstructured":"BynaS.BD\u2010CATS\u2010IO Version 00.2017.https:\/\/www.osti.gov\/biblio\/1459439"},{"key":"e_1_2_9_25_1","first-page":"59:1","volume-title":"Supercomputing","author":"Byna S","year":"2012"},{"key":"e_1_2_9_26_1","doi-asserted-by":"publisher","DOI":"10.1088\/0004\u2010637X\/765\/1\/39"},{"key":"e_1_2_9_27_1","doi-asserted-by":"crossref","unstructured":"KimY GunasekaranR ShipmanGM DillowDA ZhangZ SettlemyerBW.Workload characterization of a leadership class storage cluster. Paper presented at: 2010 5th Petascale Data Storage Workshop (PDSW) 2010. IEEE. New Orleans LA USA;2010. doi:10.1145\/2834976.2834985","DOI":"10.1109\/PDSW.2010.5668066"},{"key":"e_1_2_9_28_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-29740-3_34"},{"key":"e_1_2_9_29_1","first-page":"148","volume-title":"Icpp'12","author":"Ibtesham D","year":"2012"},{"key":"e_1_2_9_30_1","unstructured":"HueblA LeheR VayJ\u2010L et al.openPMD: a meta data standard for particle and mesh based data.2015.10.5281\/zenodo.1167843"},{"key":"e_1_2_9_31_1","unstructured":"KollerF PoeschelF GuJ HueblA.openPMD\u2010api 0.10.3: C++ & Python API for Scientific I\/O with openPMD.2019. doi:10.14278\/rodare.209"},{"key":"e_1_2_9_32_1","series-title":"Lecture Notes in Computational Science and Engineering","volume-title":"Adaptive Mesh Refinement\u00a0\u2013\u00a0Theory and Applications: Proceedings of the Chicago Workshop on Adaptive Mesh Refinement Methods, Sept. 3\u20105, 2003","author":"Plewa T","year":"2004"},{"key":"e_1_2_9_33_1","unstructured":"E3SM Project D.Energy exascale earth system model v1.3. [computer software].2019. doi:10.11578\/E3SM\/dc.20210924.5"},{"key":"e_1_2_9_34_1","doi-asserted-by":"publisher","DOI":"10.1029\/2019MS001821"},{"key":"e_1_2_9_35_1","unstructured":"Cray.Cray data warp web page.2012.http:\/\/www.cray.com\/products\/storage\/datawarp"},{"key":"e_1_2_9_36_1","doi-asserted-by":"crossref","unstructured":"LofsteadJ PolteM GibsonG et al.Six degrees of scientific data: Reading patterns for extreme scale science IO. Paper presented at: Proceedings of the 20th International Symposium on High Performance Distributed Computing HPDC'11. ACM. Association for Computing Machinery.2011; San Jose California USA: 49\u201060. doi:10.1145\/1996130.1996139","DOI":"10.1145\/1996130.1996139"},{"key":"e_1_2_9_37_1","doi-asserted-by":"crossref","unstructured":"PatwaryMMA BynaS SatishNR et al.BD\u2010CATS: big data clustering at trillion particle scale. Paper presented at: SC'15: Proceedings of the International Conference for High Performance Computing Networking Storage and Analysis. IEEE. Austin Texas USA;2015:1\u201012. doi:10.1145\/2807591.2807616","DOI":"10.1145\/2807591.2807616"},{"key":"e_1_2_9_38_1","doi-asserted-by":"crossref","unstructured":"BezJL TangH XieB et al.I\/O bottleneck detection and tuning: connecting the dots using interactive log analysis. Paper presented at: 2021 IEEE\/ACM Sixth International Parallel Data Systems Workshop (PDSW). IEEE. St. Louis MO USA;2021:15\u201022. doi:10.1109\/PDSW54622.2021.00008","DOI":"10.1109\/PDSW54622.2021.00008"}],"container-title":["Concurrency and Computation: Practice and Experience"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/cpe.8046","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/cpe.8046","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,6,25]],"date-time":"2024-06-25T01:52:27Z","timestamp":1719280347000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1002\/cpe.8046"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,4,15]]},"references-count":37,"journal-issue":{"issue":"16","published-print":{"date-parts":[[2024,7,25]]}},"alternative-id":["10.1002\/cpe.8046"],"URL":"https:\/\/doi.org\/10.1002\/cpe.8046","archive":["Portico"],"relation":{},"ISSN":["1532-0626","1532-0634"],"issn-type":[{"type":"print","value":"1532-0626"},{"type":"electronic","value":"1532-0634"}],"subject":[],"published":{"date-parts":[[2024,4,15]]},"assertion":[{"value":"2023-01-13","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-01-31","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-04-15","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e8046"}}