Generic framework for data-race-free many-particle simulations on shared memory hardware

Recently, there has been much progress in the formulation and implementation of methods for generic many-particle simulations. These models, however, typically either do not utilize shared memory hardware or do not guarantee data-race freedom for arbitrary particle dynamics. Here, we present both a abstract formal model of particle dynamics and a corresponding domain-specific programming language that can guarantee data-race freedom. The design of both the model and the language are heavily inspired by the Rust programming language that enables data-race-free general-purpose parallel computation. We also present a method of preventing deadlocks within our model by a suitable graph representation of a particle simulation. Finally, we demonstrate the practicability of our model on a number of common numerical primitives from molecular dynamics.