Dimitri Kusnezov
National Nuclear Security Administration
Friday, January 26, 2007
4:00 pm in SPL 57
The New Frontier of Computational Science: Applications to High Leverage Decisions
Abstract: Today, we are on the threshold of petaflop computing. While there are many technical challenges at this scale, this milestone is not specifically significant. Rather I believe it symbolizes the scale at which our approach to computational science - today's archetype - requires revision. We are reaching the point where raw computational power is no longer the limiting factor - rather what constrains us is our vision and our ability to ensure the accuracy of our calculational results. Computing at the petascale will open doors to discovery, and we must not be limited by the current way we approach computing. At the same time, expectations for great achievement will likely outpace our ability to effectively use these systems for innovation. This is becoming critical as these results have the potential to exert greater influence on high-consequence decisions.
[RETURN]