Molecular dynamics (MD) simulations have transformed our understanding of the nanoscale, driving breakthroughs in materials science, computational chemistry, and several other fields, including biophysics and drug design. Even on exascale supercomputers, however, runtimes are excessive for systems and timescales of scientific interest. Here, we demonstrate strong scaling of MD simulations on the Cerebras Wafer-Scale Engine. By dedicating a processor core for each simulated atom, we demonstrate a 457-fold improvement in timesteps per second versus the Frontier GPU-based Exascale platform, along with a large improvement in timesteps per unit energy. Reducing every year of runtime to less than a day unlocks currently inaccessible timescales of slow microstructure transformation processes that are critical for understanding material behavior and function.Our dataflow algorithm runs Embedded Atom Method (EAM) simulations at rates over 699k timesteps per second for problems with up to 800k atoms. This demonstrated performance is unprecedented for general-purpose processing cores.

As the US Department of Energy (DOE) computing facilities began deploying petascale systems in 2008, DOE was already setting its sights on exascale. In that year, DARPA published a report on the feasibility of reaching exascale. The report authors identified several key challenges in the pursuit of exascale including power, memory, concurrency, and resiliency. That report informed the DOE's computing strategy for reaching exascale. With the deployment of Oak Ridge National Laboratory's Frontier supercomputer, we have officially entered the exascale era. In this paper, we discuss Frontier's architecture, how it addresses those challenges, and describe some early application results from Oak Ridge Leadership Computing Facility's Center of Excellence and the Exascale Computing Project.

Billion atom molecular dynamics (MD) using quantum-accurate machine-learning Spectral Neighbor Analysis Potential (SNAP) observed long-sought high pressure BC8 phase of carbon at extreme pressure (12 Mbar) and temperature (5,000 K). 24-hour, 4650 node production simulation on OLCF Summit demonstrated an unprecedented scaling and unmatched real-world performance of SNAP MD while sampling 1 nanosecond of physical time. Efficient implementation of SNAP force kernel in LAMMPS using the Kokkos CUDA backend on NVIDIA GPUs combined with excellent strong scaling (better than 97% parallel efficiency) enabled a peak computing rate of 50.0 PFLOPs (24.9% of theoretical peak) for a 20 billion atom MD simulation on the full Summit machine (27,900 GPUs). The peak MD performance of 6.21 Matom-steps/node-s is 22.9 times greater than a previous record for quantum-accurate MD. Near perfect weak scaling of SNAP MD highlights its excellent potential to advance the frontier of quantum-accurate MD to trillion atom simulations on upcoming exascale platforms.