Computational and modeling challenges to model materials under extreme conditions

Technological innovation requires novel materials which can operate under extreme conditions of pressure, temperature, radiation damage, etc. Experiments to understand behavior of materials under such conditions are difficult and expensive, and simulations are needed to provide insight and understanding to improve desired properties. However, simulations at the atomic scale typically involve 1e5-1e7 atoms, only reaching length scales of tens of nanometers, for sub-ns time scales. New advances in both in software and hardware allow pushing the limit of simulations to reach the micron-scale comparable to some current state-of-the-art experiments. Several examples of such simulations, from astrophysics to new materials for nuclear reactors, will be presented. I will discuss challenges to push the size scales even further, including recent parallel data analysis and visualization techniques, and the use of mixed-architectures with CPUs and GPUs (Graphic Processing Units). In addition to hardware and software challenges, there are additional issues related to the models used at extreme conditions: I will present results focusing on current modeling limitations for materials at high pressures and materials under irradiation.