Last Updated: Thursday, 20 April 2017 Published: Tuesday, 18 April 2017 Print Email

From the Directors

Welcome to the new PSC Science Highlights! This first issue of our new format highlights a number of success stories in our ongoing mission to enable and support computationally dependent research in traditional and emerging fields, including the life sciences, public health, networking; to educate the next generation of researchers in using high-performance computing and big data systems (HPC); and to architect and deploy novel HPC systems.

 

The last year saw major developments in our HPC systems, furthering our efforts to better serve existing and new HPC users.The NSF-funded Phase-2 completion build of our Bridges system, accomplished on time and within budget, gives the heterogeneous, uniquely flexible Bridges increased speed, memory and storage. Quite independently, D.E. Shaw Research (DESRES) replaced its Anton 1 machine at PSC by an Anton 2, which supports much longer simulation times of much larger molecules. The only such machine available for open research, its operational funding comes from the NIH.

Our featured article is an artificial intelligence (AI) story made possible by Bridges: For the first time, an AI program has beaten the world’s best humans at an “incomplete information” game: in this case, “heads-up, no-limit Texas hold’em” poker. Libratus, a creation of researchers at the Carnegie Mellon School of Computer Science, used roughly 80 percent of Bridges’ powerful capacity to decisively defeat four of the world’s best specialists in this form of poker. The victory carries major relevance to optimizing solutions to important, real-world problems such as cybersecurity, antiterror operations, business negotiations and even medicine.

We also report on the use of Bridges in a study of the factors that affect feasibility of electric power storage for commercial users.

In addition to providing details on Anton-2, we describe a study performed on the earlier Anton system exploring the dependence of protein molecule functionality on their proper folding. The potential application range of this work includes cancer and degenerative brain diseases.

Our News Brief sectioncovers additional achievements by our Biomedical Applications Group, our STEM education efforts, research by our Network Applications Group. An important collaboration with the popular Galaxy framework and scientific gateway at Penn State is making Bridges available to genomics researchers and other scientists using Galaxy.

As always, we would like to acknowledge all our funders, especially the NSF, the NIH and the Commonwealth of Pennsylvania. We’d also like to thank our staff for the superlative work that made all these successes happen.