Earth science researchers face a major computational challenge when using AI/DL applications, but the use of GPUs in these applications has shown the potential to speed up the process. This session provided valuable insights into how researchers can optimize their AI/ML applications and improve performance through the adoption of GPU computing. The speakers for this event are Stan Posey, who provided an overview of NVIDIA support to Earth Sciences, Daniel Q. Duffy from NASA, also our IAB chair, who discussed the Hybrid Cloud Capabilities in Support of AI/ML, Jordan Alexis Caraballo-Vega from NASA Goddard, who presented an overview of AI/ML usage at Goddard, and Phil Yang, our center director, reported the GPU support to Earth Sciences Computing Testbed center IAB project. Attendees are advised to have a general understanding of AI/ML and big data in earth sciences. During the session, the various Earth science AI applications are highlighted and the maturity level of GPU-supported computing platforms such as supermicro, DGX clusters, and cloud computing are evaluated. The session concluded with a discussion on a potential hackathon at the summer ESIP meeting, aimed at further promoting the integration of GPUs in Earth science research for open science.
For more information visit, https://2023januaryesipmeeting.sched.com/event/1EwXH/utilizing-gpus-in- machine-learning-for-earth-sciences