Overview
Learn about Los Alamos National Laboratory's innovative approach to handling massive simulation data in this 39-minute conference talk. Explore how LANL tackles the challenge of managing petabyte-scale data generated from simulations, where each timestep produces enormous amounts of information requiring significant computational resources. Discover the lab's investigation into computational storage techniques aimed at reducing data volume closer to the storage source, potentially decreasing data movement requirements and analytics platform size by orders of magnitude. Follow their exploration of various approaches including row, column, object, file, and block-based methods for handling simulation output, particularly focusing on scenarios where features of interest may occupy only a fraction of the total data volume, such as tracking shockwave fronts through materials. Gain insights into LANL's collaborative efforts with partners to optimize data analytics processes and understand their strategic direction for future implementations.
Syllabus
Regional SDC Austin 24: LANL’s Journey Toward Computational Storage
Taught by
SNIAVideo