Exascale and Beyond: Challenges and Opportunities for the Materials Science and Chemistry Simulation, Modeling and Experimental Science Communities

Jack Deslippe (Hosted by Yang), Lawrence Berkeley Lab

The end of Dennard scaling (processor frequency increases) and the near end of Moore's Law (transistor area and power density) are leading to significant changes in computer architecture. As the High Performance Computing (HPC) community marches towards  "Exascale" (system with 1018 peak flops), the Top500 list of HPC systems in the world is being dominated by supercomputers employing accelerators (e.g. GPUs) and so-called many-core energy efficient processors. Many experts in the field are predicting a diversity in computer architectures exemplified by recent hardware like Google TPUs, Amazon's custom ARM processors, Intel Nervana and the possibility of Neuromorphic and Quantum Computers on the horizon.

How will all this affect the computational materials science community?

I'll discuss the important trends and postulate on the challenges and opportunities that our community will need to take advantage of or overcome to continue to push the limits of scale and fidelity in scientific predictions and data analysis. I'll discuss the material science problems being tackled on today's largest scale and being readied for the first US exascale systems in both simulation and modeling as well as data analysis from current and next generation light-sources. I'll discuss case studies of particular science applications that are combining new methods and algorithms with HPC scale to reach studies of system with unprecedented levels of size, accuracy and time.