Ambient Noise Processing With Julia
Date: 4/25/2019
Time: 04:15 PM
Room: Elliott Bay
As the total amount of global seismic noise data reaches beyond a Petabyte in the next decade, a new paradigm will be required to process large datasets. This will be particularly relevant for Distributed Acoustic Sensing (DAS) datasets going into the future, which can generate Terabytes of data a day. Here, we present a new solution for high performance seismic processing - the Julia computing language. Julia is the first dynamic, high-level language to achieve petaflop performance. Julia was designed to easily scale across CPU and GPU compute clusters. We test newly developed ambient noise cross-correlation codes written in Julia on one year of seismic data from the entire Southern California Seismic Network (SCSN).
Presenting Author: Tim Clements
Authors
Tim Clements thclements@g.harvard.edu Harvard University, Cambridge, Massachusetts, United States Presenting Author
Corresponding Author
|
Marine A Denolle mdenolle@fas.harvard.edu Harvard University, Cambridge, Massachusetts, United States |
Ellen Yu eyu@gps.caltech.edu Southern California Seismic Network, Pasadena, California, United States |
Zachary E Ross zross@gps.caltech.edu California Institute of Technology, Pasadena, California, United States |
Egill Hauksson hauksson@caltech.edu California Institute of Technology, Pasadena, California, United States |
Ambient Noise Processing With Julia
Category
Large Data Set Seismology: Strategies in Managing, Processing and Sharing Large Geophysical Data Sets