University of North Texas implements Bright Computing technology

Bright Computing, a global leader in cluster and cloud infrastructure automation software, is proud to announce that the University of North Texas (UNT) has integrated Bright Cluster Manager into its HPC environment for research support.

The University uses this 350-node cluster to optimise efficiency for researchers working on groundbreaking discoveries in a variety of domains including material science, engineering, chemistry, and physics. Bright Cluster Manager helps UNT’s University Information Technology department successfully provision, monitor and manage heterogeneous clusters.

Bright’s unified management platform allows UNT’s IT support staff to easily update and configure a variety of hardware nodes streamlining their work so that they can dedicate more time to supporting their researchers.

The heterogeneous cluster contains four kinds of Dell PowerEdge compute nodes which include: C6320 servers with two 2.4GHz Intel Xeon E5-2680 v4 fourteen-core processors, R420 servers with two 2.1GHz Intel Xeon E5-2450 eight-core processors, R720 servers with four 2.4GHz Intel Xeon E5-4640 eight-core processors, and R730 servers with two 2.4GHz Intel Xeon E5-2680 v4 fourteen-core processors and two Nvidia Tesla K80 GPUS (4,992 GPU cores/card).

UNT's HPC cluster needed an overhaul and expansion, and Bright was chosen as the solution that could best manage the older equipment along with the new expansion. The ability to manage the cluster from a single pane of glass really influenced their decision to choose Bright Cluster Manager, allowing IT staff to manage a diverse set of resources with one standard procedure which saves them time and money.

Established in 1890 and located in Denton, Texas, UNT enrolls 38,000 students and offers more than 200 degree programs at undergraduate, masters, and doctorate levels. The University is working on a diverse set of research activities that focus on materials science and engineering. Among other functions, the clusters are assisting researchers with projects like the investigation of metal fatigue in jet engines, and development of software to model basic characteristics of molecular physics. UNT has a contract with the U.S. Department of Defense to research how to make flying safer and understand physical wear factors and predictions regarding engine failure. In physics research, specifically in particle and cell physics, UNT’s researchers are looking for more accurate methods for laser beam predictability for military and civilian applications.

UNT has been using Bright Cluster Manager for seven months, and can now manage about 200 more machines. They use both the command line interface and GUI for monitoring, provisioning, and forecasting for four different classes of machines and have doubled the size of their cluster with the same amount of staff, which allows them to support scientists very efficiently.

Bright Cluster Manager enables administrators to add users, automate routine tasks, and utilise batch functionality. Using the GUI, the administrator can see all nodes in the cluster in a single pane of glass and can quickly determine if there are issues within the environment. In addition, utilising Bright Cluster Manager, users can now access basic job tracking information through the user portal. UNT has plans to use Bright’s technology to analyse workloads to forecast demand and researcher’s future needs.

Bill Wagner, CEO of Bright Computing states: ‘We are pleased that our solutions help support the researchers’ wide variety of work done in physics, chemistry, and engineering. We look forward to continuing to help UNT meet their cluster demands and research goals by simplifying their cluster monitoring and management so they can make groundbreaking discoveries.’

Other tags: 
Analysis and opinion

Robert Roe looks at research from the University of Alaska that is using HPC to change the way we look at the movement of ice sheets


Robert Roe talks to cooling experts to find out what innovation lies ahead for HPC users

Analysis and opinion