Visualization and Interpretation of Cluster Usage and User Datasets
TimeTuesday, July 306:30pm - 8:30pm
LocationCrystal Foyer and Crystal B
DescriptionIn the effort to make data-driven decisions to support end users, research computing teams gather various usage statistics about compute resource requests, runtimes and software loads. Auxiliary user data sources, such as department and domain information can be integrated to make inferences about shared features across the customer base.
This project explores various methods of visualizing and processing cluster workloads and user data. Datasets reporting the usage of the Agave cluster at Arizona State University are mined and interrogated, employing statistical and graphical algorithms for analysis.
This work complements the PEARC “Facilitation of Advanced Research Computing” technical track with an effort to better understand and support the customers of research computing systems.