Hyperwavve: A cloud-native solution for hyperscale seismic imaging on Azure

PreviousNext No AccessFirst International Meeting for Applied Geoscience & Energy Expanded AbstractsHyperwavve: A cloud-native solution for hyperscale seismic imaging on AzureAuthors: Qie ZhangGeorge IordanescuWee Hyong TokSverre Brandsberg-DahlHari Krishnan SrinivasanRanveer ChandraNavjot KukrejaGerard GormanQie ZhangMicrosoft.Search for more papers by this author, George IordanescuMicrosoft.Search for more papers by this author, Wee Hyong TokMicrosoft.Search for more papers by this author, Sverre Brandsberg-DahlMicrosoft.Search for more papers by this author, Hari Krishnan SrinivasanMicrosoft.Search for more papers by this author, Ranveer ChandraMicrosoft.Search for more papers by this author, Navjot KukrejaImperial College-London.Search for more papers by this author, and Gerard GormanImperial College-London.Search for more papers by this authorhttps://doi.org/10.1190/segam2021-3594908.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractAs cloud-computing becomes more and more popular lately, we explore its potential for hyperscale seismic imaging workloads on Azure. We introduce our cloud-native fault-tolerant solution named Hyperwavve which is based on advanced cloud technologies including Docker/Container, Kubernetes and Dask. We demonstrate a large-scale 3D FWI using 1000 VMs/nodes on Azure, where Hyperwavve uses distributed containerized processes to successfully invert for the full 3D (20x20x5 km3) overthrust velocity model. We also further validate that our Hyperwavve can distribute FWI work onto 6000 (or more) VMs/nodes concurrently. Last, we show that our Python-based FWI runs on both Azure CPUs and GPUs including various architectures.Keywords: full-waveform inversion, cloud computing, distributed systems, parallel, reverse time migrationPermalink: https://doi.org/10.1190/segam2021-3594908.1FiguresReferencesRelatedDetails First International Meeting for Applied Geoscience & Energy Expanded AbstractsISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2021 Pages: 3561 publication data© 2021 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished: 01 Sep 2021 CITATION INFORMATION Qie Zhang, George Iordanescu, Wee Hyong Tok, Sverre Brandsberg-Dahl, Hari Krishnan Srinivasan, Ranveer Chandra, Navjot Kukreja, and Gerard Gorman, (2021), "Hyperwavve: A cloud-native solution for hyperscale seismic imaging on Azure," SEG Technical Program Expanded Abstracts : 782-786. https://doi.org/10.1190/segam2021-3594908.1 Plain-Language Summary Keywordsfull-waveform inversioncloud computingdistributed systemsparallelreverse time migrationPDF DownloadLoading ...