PNNL

Abstract

At what point does a faster computer fundamentally change how you or your laboratory performs research? Organizations and scientists who depend on computers to reconstruct instrument data or who utilize simulations as part of their workflow most certainly have a sense of how faster computers can benefit their work. How about you – if the price were right would a 100x or faster computer fundamentally change your research? Dual and quad core vendors are actively hawking the benefits of their wares. In terms of changing the landscape of scientific computing, I believe a 2x-4x performance increase is an “incremental” change. The computer will do more and the scientist will wait less. In other words, existing workflows benefit as the computer becomes less of a bottleneck. Please don’t get me wrong, these machines are nice and I will be actively looking to upgrade my workstation when the time comes, but the performance increase is incremental.