Skip to main content
NASA Advanced Supercomputing (NAS) Division World-class modeling, simulation, and supercomputing in support of the NASA mission

Rapid Analysis and Management of Large Spatiotemporal Data

AMS Seminar Series

Speaker: Ben O’Neill, RNET Technologies and Rohit Deshmukh, Florida Atlantic University

Presentation

Seminar Slide Deck (PDF-2MB)

Abstract

NASA's vast collection of simulated, experimental, and observational datasets has paved the way for groundbreaking advancements in addressing real-world physical challenges. Yet, the sheer size of high-fidelity data generated presents challenges in terms of data transfer, computer resources, and data volume, particularly for downstream users lacking direct access to high performance computing resources.

In this talk, we introduce a lossy reduce-then-analyze strategy that enables efficient data storage, transfer, and analysis. Our data reduction method, called the Data-informed Local Subspaces (DLS), uses data-driven feature learning and data-enriched finite element methods to transform high fidelity data into a reduced format that promotes rapid data analytics. Results demonstrating 20-100x data reduction for canonical fluid dynamics datasets, and 10-80x for earth science datasets will be discussed. An order of magnitude speedup in vorticity calculations and multiple orders of magnitude improvement in signal processing analyses will also be presented. Finally, we will discuss a pathway for how our tools can be integrated into NASAs existing aerospace and earth science workflows.

This work is made possible through the support of a NASA Phase I Small Business Innovation Research (SBIR) award.

Biography

Photo of Ben O’Neill
Ben O’Neill is a senior research scientist at RNET Technologies Inc. in Dayton, Ohio. Ben received his Ph.D. in Applied Mathematics from the University of Colorado at Boulder with a specialization in high performance computing and parallel time integration. Since joining RNET, Ben has been the driving force behind RNETs line of HPC software usability tools, including SolverSelector – a machine learning tool for optimal linear solver selection, VnVLabs – a platform for scientific simulation in the Cloud, and MLDC - our DLS based Rapid Analytics Platform.



Photo of Rohit Deshmukh
Rohit Deshmukh is an Assistant Professor at Florida Atlantic University (FAU). Rohit obtained his Ph.D. in Aerospace Engineering from The Ohio State University in 2016, with a focus on applying model order reduction to incompressible turbulent flows. Rohit has extensive research experience in computational mechanics in the areas of model reduction, data-driven-physics-based model development, and coupled fluid-structural-thermal simulation and analysis.