Researchers from Argonne and Texas A&M University will present a technical paper on a new data-driven diagnostic tool called Gauge that can be used to explore the latent space of supercomputing job features, understand behaviors of clusters of jobs and identify and assess I/O bottlenecks.