In Summer 2023, through the Sustainable Research Pathways (SRP)  Program,  I was selected to contribute to a project in the Mathematics and Computer Science (MCS) Division. There, I worked with Dr. Kevin A. Brown to develop visual analytics tools to explore best ways to  optimize the performance of high-performance computing (HPC) systems.  To achieve this, I  enhanced a Node.js app for analyzing and visualizing large, complex simulation datasets. My work made it feasible to  aggregate network traffic by group, rank, or job, allowing us to observe causality between entities and assess variance and saturation of metrics to deduce the trends in topological groupings.

We employed our application to conduct a focused analysis and visualization  of  the MiniFE mini-app to emulate and understand real scientific workloads. These tools and methods can help guide configuration and performance tuning for cutting-edge HPC systems at Argonne like Aurora and Polaris. It was energizing to contribute to efforts that guide the building of the fastest TOP500 supercomputers by improving analytic workflows and simulation-based tuning.