Hello,
Please join the eScience Institute Wednesday, May 1, 4:00 pm in EEB-303. Refreshments will be provided.
Jeff Gardner (UW Physics)
Jeff Gardner is Director of Research for Physical Sciences at the eScience Institute, Affiliate Assistant Professor in the Physics and Astronomy departments, and Visiting Faculty at Google, Inc. Jeff received his PhD in Astronomy from UW in 2000. In 2003, he become a Sr. Research Scientist at the Pittsburgh Supercomputing Center, where he participated in the deployment of the NSF TeraGrid (Extensible Terascale Facility; ETF), which became the largest open platform for scientific computing in the world. His research has focused on the overcoming the challenges of analyzing extremely large scientific datasets using a variety of approaches, including scalable DBMSs, MapReduce, as well as domain-specific libraries. He is also actively involved in building the next generation of computational astrophysics codes capable of sustaining a petaflop (1 thousand trillion mathematical operations per second) and generating petabytes of data.
Simulating the Universe on Google’s Exacycle Platform
The Large Synoptic Survey Telescope (LSST; http://www.lsst.org ) is one of the most ambitious astrophysical research programs ever undertaken. From the 9,000 ft Cerro Pachon peak in Northern Chile, the LSST’s 3.2 Gigapixel camera will repeatedly survey the southern sky, taking one image every 15 seconds, generating tens of petabytes of data every year. The images and catalogs from the LSST have the potential to transform both our understanding of the universe and the way that we undertake science. As part of the implementation phase of this project, the LSST collaboration has undertaken a formidable program to simulate the flow of data from the telescope. The image simulator traces individual photons of light from stars, galaxies, asteroids, through the earth’s atmosphere, the telescope optics, and onto the detector. These simulations are used to optimize how the LSST surveys the sky, to develop the analytics required to understand how the universe forms and evolves, and to determine how astronomers (and the public as a whole) will scale science to data sets that will exceed a hundred petabytes in size. For over a year now, Google has given LSST access to their Exacycle platform in order to perform these simulations (http://research.google.com/university/exacycle_program.html), reducing the time required to simulate one night of LSST observing, roughly 5 million images, from 3 months down to a few days. This rapid turnaround enables the LSST engineering teams to test new designs and new algorithms with unprecedented precision, which will ultimately lead to bigger and better science.
Upcoming Seminars:
* May 13, 4 PM (EE303)
Fernando Perez (Berkeley)
TBD
* May 22, 4 PM (EE303)
Joe Hellerstein (Berkeley)
Why Computer Scientists Should and Can Learn Computer Science
