Diffusion Monte Carlo: Exponential scaling of computational cost for large systems

Abstract

The computational cost of a Monte Carlo algorithm can only be meaningfully discussed when taking into account the magnitude of the resulting statistical error. Aiming for a fixed error per particle, we study the scaling behavior of the diffusion Monte Carlo method for large quantum systems. We identify the correlation within the population of walkers as the dominant scaling factor for large systems. While this factor is negligible for small and medium sized systems that are typically studied, it ultimately shows exponential scaling. The scaling factor can be estimated straightforwardly for each specific system and we find that is typically only becomes relevant for systems containing more than several hundred atoms.