John Grosh is the Director of the Center for Applied Scientific Computing at LLNL. He covered elements of LLNL’s petascale computing strategy, including Sequoia.
A part of his talk that I was particularly interested in is the work going on at LLNL dealing with architectures to process very large volumes of data. They are studying models from the business intelligence world and even prototyping some of those systems to understand how they apply in a scientific environment.
Be sure to check out my article in the upcoming edition of HPCwire where I’ll talk more about some of that work as well as the Pervasive DataRush solution.