Adaptable Parallel Acceleration Strategy for Dynamic Monte Carlo Simulations of Polymerization with Microscopic Resolution

Properties of polymer products are determined by their microscopic structures. Dynamic Monte Carlo (DMC) simulation is a powerful tool to capture detailed polymer microstructures. However, the heavy computational burden significantly limits the broad application of DMC in practice. In this work, a sub-box parallel strategy is proposed to accelerate the DMC simulation of polymerization processes with accuracy assurance. Different parallelizable task granularities are derived for different kinetic reaction mechanisms. The integration of reaction mechanism and hardware architecture is fully addressed by proposing the implementation of the simulations on a multicore processor platform or a graphics processing unit platform according to parallelizable task granularity. Key parallelization questions including random number generation, data structure, and communication strategy are purposely answered for different scenarios. Five case studies with kinetic mechanisms of increasing complexity, including linear and branching polymerizations, are presented to show the efficiency, accuracy, and reliability of the proposed parallelization strategy.