Supercomputing 2018

The Institute for CyberScience (ICS) will be hosting a booth at Supercomputing 2018, and we are looking forward to talking with you about everything we do to enable and support HPC research for Penn State! Join us at booth #4319 for presentations covering our HPC operations, researcher engagement, NIST certification, and the future of HPC research infrastructure at Penn State. And don’t forget to get a selfie with our favorite Penn Stater, the Nittany Lion mascot!

Booth Talks on Wednesday, November 14

11:00 a.m.: “Measuring the Impact of Computational Research Facilities”

Presented by Wayne Figurelle, Assistant Director, ICS

ICS put together a report showing the impact of HPC facilities on academic productivity at Penn State. Wayne will discuss the importance of metrics, how ICS chose the right metrics, and give a brief overview of some of the report’s findings.

1:00 p.m.: “The Penn State ICS High-Performance Research Cloud”

Presented by Chuck Pavloski, Chief Architect, ICS

The Penn State Institute for CyberScience is in the process of developing and deploying an on- and off-premises research computing cloud to compliment the HPC and HTC bare-metal offerings in place. Chuck will explain the exciting features and future opportunities this provides for applied computational research at Penn State.

2:00 p.m.: “Obtaining NIST certification for HPC environments”

Presented by Ryan Gilmore, Technical Director, ICS

Ryan will discuss the challenges, considerations, and overall approach that ICS took while designing for a NIST compliant system. He will highlight design decisions and lessons learned.

3:00 p.m.: “Deploying CVMFS on a Shared Supercomputer for Happy Users and Upstream Admins”

Presented by Adam Focht, Implementation Team Lead, ICS

Adam will discuss considerations such as minimizing hits to the remote file system, minimizing data retrieval latency, and maximizing availability. ICS used Puppet to deploy the configuration, utilized an alien cache on a shared NFS file system, and placed a Squid proxy via Docker to handle the vast majority of remote requests.

4:30 p.m.: “Researcher Engagement Initiatives”

Presented by Derek Leydig, Client Engagement Manager, ICS

5:30 p.m.:  “Controlling Operations”

Presented by Gary Skouson, Operations Team Lead, ICS

Operating HPC systems (or many other complex systems) doesn’t automatically just work out.  You need to develop a framework for operations and design to that framework.  Understanding operational constraints and designing or not for those constraints will impact the functionality, availability and flexibility of your HPC system.

Open positions

View open ICS staff positions.

For open co-hired faculty positions (shared with academic colleges), visit the following links: