Colloquium Details
Descriptive Data and Productive Programming: Addressing Analysis Challenges at Extreme Scales
| Author: | Jeremy Meredith Oak Ridge National Laboratory |
|---|---|
| Date: | January 16, 2014 |
| Time: | 15:30 |
| Location: | 220 Deschutes |
| Host: | Hank Childs |
Abstract
Analysis and visualization of the data generated by scientific simulation codes is a key step in enabling science from computation. However, a number of obstacles lie along the current hardware and software paths to exascale scientific discovery. For example, only advanced parallelism techniques will take full advantage of the unprecedented scale of the upcoming machines; the limited memory environment, particularly in the context of in situ analysis, will require efficiency of both algorithms and infrastructure; and the advancement of simulation codes with complex data models requires commensurate support in analysis libraries. This talk will explore these challenges and recent research intended to address them.
Biography
Jeremy Meredith is a computer scientist with the Future Technologies Group in the Computer Science and Mathematics Division of Oak Ridge National Laboratory. He researches next-generation computing architectures, including evaluating GPUs for general purpose computation. He is a founding member of the development team for the VisIt scientific visualization application. Before coming to ORNL, he worked for Lawrence Livermore National Laboratory. He received his MS in Computer Science from Stanford University and his BS from the University of Illinois at Urbana-Champaign.
Jeremy has received many awards, including a 2008 ACM Gordon Bell Prize, a 2005 R&D 100 Award, and DOE/NNSA Defense Program's Awards of Excellence in 1999 and 2002.