How Do I Rate This?
The blue stars show the average user rating for this item. To add your own rating, move your cursor over the stars to highlight them in gold, and click to show your rating. One star highlighted is the lowest rating, all five is the highest. Once you have rated an item, your rating is added to the average.
Multiscale Dataflow Computing
Date: Tuesday, April 23, 2013 Time: 1:00pm Location: 321-B20 Speaker: Oskar Mencer
Solar System Exploration Directorate (4X) and NSTA (8X) Seminar
Complexity of computation is a function of the underlying representation. We are extending this basic concept to consider representation of computational problems on the application level, the model level, the architecture level, arithmetic level and gate level of computation. In particular, the first step is to consider and optimize the discretization of a problem in time, space and value. Discretization of value is particularly painful, both in Physics where atomic discretization ruins many nice theories, and in computation, where most people just use IEEE double precision floating point so they don’t have to worry about details, until they do. Multiscale Dataflow Computing provides a process by which one can optimize the discretization of time, space and value based on a particular underlying computer architecture, and in fact, iterate the molding of the computer architecture and the discretization of the computational challenge.
Commercial Maxeler systems in production have been reported to achieve 10-50x faster computation per cubic foot and per Watt, resulting in less nodes per computation and therefore exponentially improved reliability and resiliency.
Results published by customers include financial modeling (American Finance Technology Award for most cutting edge technology, 2011), 3D finite difference in the Oil&Gas industry (see Society of Exploration Geophysicists meetings and reports), weather modeling (reducing time to compute a Local Area Model – LAM from 2 hours to 2 minutes) and even sparse matrix solvers (finite elements) which cannot be parallelized, running 20-40x faster.
Prior to founding Maxeler, Oskar Mencer was Member of Technical Staff at the Computing Sciences Center at Bell Labs in Murray Hill, leading the effort in "Stream Computing". He joined Bell Labs after receiving a PhD from Stanford University. Besides driving Maximum Performance Computing (MPC) at Maxeler, Oskar is Consulting Professor in Geophysics at Stanford University and he is also affiliated with the Computing Department at Imperial College London, having received two Best Paper Awards, an Imperial College Research Excellence Award in 2007 and a Special Award from Com.sult in 2012 for "revolutionizing the world of computers."
For further information, please contact: Leon Alkalai (x45988), Larry Bergman (x35314)