SeminarSeptember 30, 2016

3:00 PM PFNAC 013

The Digistar Planetarium as an Example of Advancing Science with Graphics Processing

Units

Loren SpencerUM Physics Major

In our changing world, the need for a scientifically knowledgeable public is increasing along with the need to make good policy decisions. In addition, the realities of science are changing. From the COBE satellite, launched in 1989, to the Large Hadron Collider, currently producing large datasets, and on into the future, the need for analysis of “big data” in science continues to increase - substantially. Both the presentation of ideas and the analysis of data are benefiting from new capabilities in parallel computing which, to a large extent, have been driven by the computer gaming industry. Hardware developed for this highly competitive, and often lucrative, gaming market is being adapted to meet the needs of scientists and educators on an increasingly cost effective basis. In terms of MegaFLOP capabilities (millions of floating point operations) computer expansion cards are available for a few hundred dollars that rival super-computers only available to military, large corporations and universities a decade ago. This technology is scalable and inexpensive; a powerful machine can be built for a few thousands of dollars.

These advances in parallel processing (hardware) and parallel programming (software) are being used at the University of Montana in Astronomy course work and public outreach to generate interest among students and the general populous in the advancement of science and STEM education (Science, Technology, Engineering and Mathematics). Cultural education and awareness are also being developed as an ongoing endeavor, seeking to be as inclusive of as many cultures as possible.

This presentation will start with an architectural overview to facilitate insight into the nature, capabilities and limitations of this technology. Second, an overview of the software interface provided by the Digistar system and its use in demonstrating Astronomical ideas is presented. Finally, the discussion will address possible applications of this technology to turn data into knowledge.