{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T03:41:36Z","timestamp":1783568496484,"version":"3.55.0"},"publisher-location":"New York, NY, USA","reference-count":18,"publisher":"ACM","license":[{"start":{"date-parts":[[2023,11,12]],"date-time":"2023-11-12T00:00:00Z","timestamp":1699747200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"DOI":"10.13039\/501100006374","name":"U.S. Department of Energy","doi-asserted-by":"publisher","award":["DE-AC05-00OR22725"],"award-info":[{"award-number":["DE-AC05-00OR22725"]}],"id":[{"id":"10.13039\/501100006374","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2023,11,12]]},"DOI":"10.1145\/3624062.3624134","type":"proceedings-article","created":{"date-parts":[[2023,11,10]],"date-time":"2023-11-10T13:53:39Z","timestamp":1699624419000},"page":"619-626","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Experiences Detecting Defective Hardware in Exascale Supercomputers"],"prefix":"10.1145","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3001-4414","authenticated-orcid":false,"given":"Nick","family":"Hagerty","sequence":"first","affiliation":[{"name":"Oak Ridge National Laboratory, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3887-7601","authenticated-orcid":false,"given":"Jordan","family":"Webb","sequence":"additional","affiliation":[{"name":"Oak Ridge National Laboratory, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4333-4145","authenticated-orcid":false,"given":"Veronica","family":"Melesse Vergara","sequence":"additional","affiliation":[{"name":"Oak Ridge National Laboratory, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4149-0638","authenticated-orcid":false,"given":"Matt","family":"Ezell","sequence":"additional","affiliation":[{"name":"Oak Ridge National Laboratory, United States of America"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2023,11,12]]},"reference":[{"key":"e_1_3_2_2_1_1","volume-title":"TOP500 List","year":"2022","unstructured":"2022. TOP500 List June 2022. https:\/\/www.top500.org\/lists\/top500\/2022\/06\/."},{"key":"e_1_3_2_2_2_1","unstructured":"[2] 2023. https:\/\/www.ornl.gov\/news\/frontier-supercomputer-debuts-worlds-fastest-breaking-exascale-barrier."},{"key":"e_1_3_2_2_3_1","unstructured":"2023. Frontier User Guide. https:\/\/docs.olcf.ornl.gov\/systems\/frontier_user_guide.html."},{"key":"e_1_3_2_2_4_1","unstructured":"2023. Grafana Open-Source. https:\/\/grafana.com\/oss\/grafana\/."},{"key":"e_1_3_2_2_5_1","unstructured":"2023. Hardware\/Hybrid Accelerated Cosmology Code. https:\/\/cpac.hep.anl.gov\/projects\/hacc\/."},{"key":"e_1_3_2_2_6_1","unstructured":"2023. High-Performance Linpack Benchmark. http:\/\/icl.cs.utk.edu\/hpl\/index.html."},{"key":"e_1_3_2_2_7_1","unstructured":"2023. InfluxDB. https:\/\/www.influxdata.com\/products\/influxdb\/."},{"key":"e_1_3_2_2_8_1","unstructured":"2023. OLCF Test Harness Examples. https:\/\/github.com\/olcf\/olcf-test-harness-examples. https:\/\/github.com\/olcf\/olcf-test-harness-examples"},{"key":"e_1_3_2_2_9_1","volume-title":"TOP500 List","year":"2023","unstructured":"2023. TOP500 List June 2023. https:\/\/www.top500.org\/lists\/top500\/2023\/06\/."},{"key":"e_1_3_2_2_10_1","unstructured":"Ulrike\u00a0Yang et al. 2017. AMG. https:\/\/github.com\/LLNL\/AMG. https:\/\/github.com\/LLNL\/AMG"},{"key":"e_1_3_2_2_11_1","volume-title":"Frontier node health checking and state management","author":"Ezell Matt","year":"2023","unstructured":"Matt Ezell. 2023. Frontier node health checking and state management. Cray User Group Proceedings (2023)."},{"key":"e_1_3_2_2_12_1","volume-title":"rocHPL. https:\/\/github.com\/ROCmSoftwarePlatform\/rocHPL. Retrieved","author":"Advanced Micro\u00a0Devices Inc. 2023.","year":"2023","unstructured":"Advanced Micro\u00a0Devices Inc. 2023. rocHPL. https:\/\/github.com\/ROCmSoftwarePlatform\/rocHPL. Retrieved August 4, 2023 from https:\/\/github.com\/ROCmSoftwarePlatform\/rocHPL"},{"key":"e_1_3_2_2_13_1","volume-title":"rocPRIM. https:\/\/github.com\/ROCmSoftwarePlatform\/rocPRIM. Retrieved","author":"Advanced Micro\u00a0Devices Inc. 2023.","year":"2023","unstructured":"Advanced Micro\u00a0Devices Inc. 2023. rocPRIM. https:\/\/github.com\/ROCmSoftwarePlatform\/rocPRIM. Retrieved August 4, 2023 from https:\/\/github.com\/ROCmSoftwarePlatform\/rocPRIM"},{"key":"e_1_3_2_2_14_1","volume-title":"Towards Acceptance Testing at the Exascale Frontier","author":"Vergara Larrea Veronica","year":"2020","unstructured":"Veronica G.\u00a0Vergara Larrea, Michael\u00a0J. Brim, Arnold Tharrington, Reuben Budiardja, and Wayne Joubert. 2020. Towards Acceptance Testing at the Exascale Frontier. Cray User Group Proceedings (2020). https:\/\/cug.org\/proceedings\/cug2020_proceedings\/includes\/files\/spec111s1.pdf"},{"key":"e_1_3_2_2_15_1","unstructured":"Arnold\u00a0N. Tharrington. 2015. NCCS Regression Test Harness Version 00. https:\/\/www.osti.gov\/\/servlets\/purl\/1232564"},{"key":"e_1_3_2_2_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cpc.2021.108171"},{"key":"e_1_3_2_2_17_1","doi-asserted-by":"publisher","DOI":"10.1109\/COGSIMA.2015.7108181"},{"key":"e_1_3_2_2_18_1","volume-title":"Balancing Workloads in More Ways than One","author":"Melesse Veragara Veronica","year":"2023","unstructured":"Veronica G.\u00a0Melesse Veragara, Paul Peltz, Nick Hagerty, Christopher Zimmer, Reuben Budiardja, Dan Dietz, Thomas Papatheodore, Christopher Coffman, and Benton Sparks. 2023. Balancing Workloads in More Ways than One. Cray User Group Proceedings (2023)."}],"event":{"name":"SC-W 2023: Workshops of The International Conference on High Performance Computing, Network, Storage, and Analysis","location":"Denver CO USA","acronym":"SC-W 2023"},"container-title":["Proceedings of the SC '23 Workshops of the International Conference on High Performance Computing, Network, Storage, and Analysis"],"original-title":[],"link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3624062.3624134","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3624062.3624134","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,8,21]],"date-time":"2025-08-21T03:00:59Z","timestamp":1755745259000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3624062.3624134"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,11,12]]},"references-count":18,"alternative-id":["10.1145\/3624062.3624134","10.1145\/3624062"],"URL":"https:\/\/doi.org\/10.1145\/3624062.3624134","relation":{},"subject":[],"published":{"date-parts":[[2023,11,12]]},"assertion":[{"value":"2023-11-12","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}