Utilities To Execute Pipelines (UTEP)

Utilities To Execute Pipelines (UTEP) Diego Juarez†, Jorge Munoz The University of Texas at El Paso †[email protected] Abstract Methodology Results from Simulation Tools Deployed with UTEP Objectives Python framework that automates the creation, submission, and execution of large-scale simulation datasets, reducing human error and saving time when managing hundreds or thousands of jobs. Deployed on the Perlmutter supercomputer at NERSC, it has proven efficient, accurate, and adaptable to other computing platforms. Statement of the Problem Input parameter settings Organizing experiment Creating auxiliary files Deploying simulations Analyzing results 1. Define experiment parameters. 2. Construct a hierarchically organized directory structure. 3. Generate supporting files. 4. Perform simulations. 5. Data visualization and data collection. oEfficiently manage big data on supercomputers, saving time and storage. oMinimize human error in simulation workflows. oAdaptable to many software packages. Embarrassingly parallel simulations. Manual job + self-management = time-consuming + error prone. UTEP automates and organize pipelines and analyzes Harmonic Ensemble Lattice Dynamics Temperature-Dependent Harmonic Model FerroX MagneX Stability map for BCC crystal using MOGA References: