A. High performance computing in the simulation sciences
Detailed numerical simulations are becoming ever more important in the natural sciences and benefit dramatically from direct collaboration between Mathematicians, Computer Scientists, Physicists, Chemists etc., in order to produce validated, reliable and effective simulations. The latter also involves a thorough understanding of current and anticipated future high-performance computing (HPC) topics. These topics, beside the efficient use of modern distributed many-core computer architectures, also include large-scale scientific data management, the development of parallel distributed computing environments, and also ahead of the curve experimentation with emerging HPC trends such as hardware accelerator (e.g., using systems based on graphics processors) based computing. In many cases, efficient use of these technologies requires not a mere re-writing of parts of the simulation computer codes, but necessitates re-thinking of the mathematical approaches taken and of the algorithms of the simulations.