Optimizing the Efficiency of the NCAR-Wyoming Supercomputing Facility

08/01/2014 - 10:55am
Mesa Lab Main Seminar Room

Theophile Nsengimana portrait

Theophile Nsengimana, SIParCS Intern
(Philander Smith College)

The proper analysis of the bulk data generated by mechanical and electrical sensors spread across the NCAR-Wyoming supercomputing center can provide invaluable insights about the facility. They can for instance assist us to distinguish the facility’s normal working condition from aberrant; which can further help us to predict the behavior of the facility in addition to providing appropriate, well-timed maintenance. As is typical with any significant sensor based data collection, a certain amount of quality control is required. Additionally the accessibility of the data is frequently an issue. The efforts of this project strove to automate this processing using Python. Additionally the project explores the usefulness of an open archive tool developed at the University of California Berkeley, Simple Measurement and Actuation Profile tool (sMAP).  

Video Presentation