SCOPED Update: A Cloud and HPC Software Platform for Computational Seismology
Description:
The goal of the SCOPED (Seismic COmputational Platform for Empowering Discovery) project is to develop a cyberinfrastructure that enables hybrid model-data research in seismology. The open-source software development will make it easier for researchers to connect and collaborate via common platforms like GitHub and containers like Docker. Seismological research is typically either data-driven or model-driven. Data-driven workflows are well-suited for cloud infrastructure, while model-driven workflows are best suited for HPC. The SCOPED platform includes tools for data discovery, theoretical seismology, wavefield simulations, and inversion workflows for velocity Earth models and earthquake source characterization. The platform containerizes our seismological software to run on both HPC and cloud platforms. Tutorials are also included to guide new users on how to run the software. Additionally, the SCOPED Gateway is available as a preliminary implementation, which enables these containers as services over the web.
In this SCOPED update, we present several advances. First, we demonstrate using the SCOPED gateway for HPC simulations of seismic wavefields and their inversion using MTUQ and SPECFEM. In addition, we will discuss the current and future I/O problems and potential remedies for large-scale full-waveform inversions on HPC systems. Second, we demonstrate cloud-native workflows on Amazon WebServes (AWS). We demonstrate how to do ambient noise seismology with NoisePy, deep-learning seismology using SeisBench and QuakeFlow, and unsupervised ML for source characterization and discovery. We demonstrate how to use AWS tools and Infrastructure-as-a-service tools such as Coiled to abstract cloud systems for easier access. Finally, we show how SCOPED ways to bridge cloud and HPC using the SeismoDB data lake.
Session: Leveraging Cutting-Edge Cyberinfrastructure for Large Scale Data Analysis and Education [Poster Session]
Type: Poster
Date: 5/2/2024
Presentation Time: 08:00 AM (local time)
Presenting Author: Marine
Student Presenter: No
Invited Presentation:
Authors
Marine Denolle Presenting Author Corresponding Author mdenolle@uw.edu University of Washington |
Carl Tape ctape@alaska.edu University of Alaska Fairbanks |
Yinzhi Wang iwang@tacc.utexas.edu University of Texas at Austin |
Ebru Bozdag bozdag@mines.edu Colorado School of Mines |
Felix Waldhauser felixw@ldeo.columbia.edu Columbia University |
Eric Beaucé ebeauce@ldeo.columbia.edu Columbia University |
Bryant Chow bhchow@alaska.edu University of Alaska Fairbanks |
Kuan-Fu Feng kffeng@uw.edu University of Washington |
Nathan Freeman nfreeman@tacc.utexas.edu University of Texas at Austin |
Carlos Garcia Jurado Suarez carlosg@uw.edu University of Washington, Seattle, Washington, United States |
Masaru Nagaso masaru.nagaso@mines.edu Colorado School of Mines, Golden, Colorado, United States |
Yiyu Ni niyiyu@uw.edu University of Washington, Seattle, Washington, United States |
Steven Oh oh.jae.steven@utexas.edu University of Texas at Austin, Austin, Texas, United States |
Ridvan Orsvuran rorsvuran@mines.edu Colorado School of Mines, Golden, Colorado, United States |
Theresa Sawi tsawi@ldeo.columbia.edu Columbia University, Palisades, New York, United States |
Joe Stubbs jstubbs@tacc.utexas.edu University of Texas at Austin, Austin, Texas, United States |
Jackson Thetford j.thetford@utexas.edu University of Texas at Austin, Austin, Texas, United States |
Julien Thurin jthurin@alaska.edu University of Alaska Fairbanks, Fairbanks, Alaska, United States |
Chenxiao Wang iwang@tacc.utexas.edu University of Texas at Austin, Austin, Colorado, United States |
Kaiwen Wang kw2988@ldeo.columbia.edu Columbia University, Palisades, New York, United States |
SCOPED Update: A Cloud and HPC Software Platform for Computational Seismology
Category
Leveraging Cutting-Edge Cyberinfrastructure for Large Scale Data Analysis and Education