CellListMap.jl: Efficient and customizable cell list implementation for calculation of pairwise particle properties within a cutoff

N-body simulations and trajectory analysis rely on the calculation of attributes that depend on pairwise particle distances within a cutoff. Interparticle potential energies, forces, distribution functions, neighbor lists, and distance-dependent distributions, for example, must be calculated. Cell lists are widely used to avoid computing distances outside the cutoff. However, efficient cell list implementations are difficult to customize. Here, we provide a fast and parallel implementation of cell lists in Julia that allows the mapping of custom functions dependent on particle positions in 2 or 3 dimensions. Arbitrary periodic boundary conditions are supported. Automatic differentiation and unit propagation can be used. The implementation provides a framework for the development of new analysis tools and simulations with custom potentials. The performance of resulting computations is comparable to state-of-the-art implementations of neighbor list algorithms and cell lists, available in specialized software. Examples are provided for the computation of potential energies, forces, distribution of pairwise velocities, neighbor lists and other typical calculations in molecular and astrophysical simulations. The Julia package is freely available at http://m3g.github.io/CellListMap.jl. Interfacing with Python and R with minimal overhead is possible.

N体模拟（N-body simulations）与轨迹分析依赖于基于截断半径内粒子对间距计算各类属性。例如粒子间势能、作用力、分布函数、邻接列表（neighbor lists）以及依赖于间距的分布等，均需进行计算。细胞列表（cell lists）被广泛用于规避对截断半径外的间距进行计算，但高效的细胞列表实现往往难以自定义适配需求。本研究提供了Julia语言中一种快速且并行的细胞列表实现，其支持对二维或三维空间中依赖粒子位置的自定义函数进行映射。该实现支持任意形式的周期性边界条件（periodic boundary conditions），同时可使用自动微分（automatic differentiation）与单位传递功能。此实现为开发基于自定义势能的新型分析工具与模拟程序提供了框架支撑。其计算性能可与专用软件中现有最先进的邻接列表算法与细胞列表实现相媲美。文中提供了势能计算、作用力计算、粒子对速度分布、邻接列表以及分子与天体物理模拟中其他典型计算任务的示例。该Julia软件包可在http://m3g.github.io/CellListMap.jl免费获取，同时可通过极低的开销实现与Python及R语言的交互。