An adaptive particle-mesh gravity solver for ENZO

TL;DR

This paper introduces an adaptive particle-mesh algorithm for solving the Poisson equation in grid-based hydrodynamics, demonstrating improved accuracy and reduced noise over traditional methods within the Enzo astrophysical simulation code.

Contribution

The paper presents a novel adaptive particle-mesh solver integrated into Enzo, enhancing accuracy for point masses and nested grid configurations compared to existing multigrid methods.

Findings

More accurate for point mass distributions

Less noisy than multigrid solver

Effective in two-body problem simulations

Abstract

We describe and implement an adaptive particle-mesh algorithm to solve the Poisson equation for grid-based hydrodynamics codes with nested grids. The algorithm is implemented and extensively tested within the astrophysical code Enzo against the multigrid solver available by default. We find that while both algorithms show similar accuracy for smooth mass distributions, the adaptive particle-mesh algorithm is more accurate for the case of point masses, and is generally less noisy. We also demonstrate that the two-body problem can be solved accurately in a configuration with nested grids. In addition, we discuss the effect of subcycling, and demonstrate that evolving all the levels with the same timestep yields even greater precision.