Particle-Mesh code for cosmological simulations

TL;DR

The paper presents a public Particle-Mesh N-body code for cosmological simulations, emphasizing a fast, simple method that scales to large grids while accepting moderate resolution. It provides an end-to-end toolkit: initial-condition generation via Zeldovich approximation and Boltzmann-based $P(k)$ fits, FFT-based Poisson solving, CIC density assignment, leapfrog integration, and a Bound-Density-Maxima halo finder. The package supports diverse cosmologies (open/flat/closed, with/without $\Lambda$, varying $H_0$, hot neutrinos, and baryon content) and enables detailed post-processing (power spectra, density distributions, and halo catalogs). Test runs and documented examples accompany the release, with explicit guidance on compilation, execution, and data formats, making it a practical resource for the astronomical community, albeit without guaranteed results.

Abstract

Particle-Mesh (PM) codes are still very useful tools for testing predictions of cosmological models in cases when extra high resolution is not very important. We release for public use a cosmological PM N-body code. We provide a complete package of routines needed to set initial conditions, to run the code, and to analyze the results. The package allows you to simulate models with numerous combinations of parameters: open/flat/closed background, with or without the cosmological constant, different values of the Hubble constant, with or without hot neutrinos, tilted or non-tilted initial spectra, different amount of baryons. Routines are included to measure the power spectrum and the density distribution function in your simulations, and a bound-density-maxima code for halo finding. We also provide results of test runs. A simulation with 256^3 mesh and 128^3 particles can be done in a couple of days on a typical workstation (70Mb of RAM are needed). To run simulations with 800^3 mesh and 256^3 particles one needs a computer with 1Gb memory and 1Gb disk space. The code has been successfully tested on an HP workstation and on a Sun workstation running Solaris. Most of the files (not tests) can be obtained from ftp://astro.nmsu.edu/pub/aklypin/PMCODE The package can be downloaded from http://astro.nmsu.edu/~aklypin/PM/pmcode/index.html We provide this tool as a service to the astronomical community, but we cannot guarantee results or publications.

Figures (3)

• Figure 1: (Top) Errors of the fits eq.(\ref{['EqFit']}) for the CDM models ($\Omega_0=1$) with a Hubble constant $H=50$km/s/Mpc. Errors at the $\sim$ 2% level at $k\sim 3h~{\rm Mpc}^{-1}$ and at $k\sim 30h~{\rm Mpc}^{-1}$ are due to a small mismatch in approximations used at high wavenumbers. The fits smooth out the jumps and, thus, provide better approximations to the real power spectra at these large wavenumbers. The waves around $k\sim 0.1h~{\rm Mpc}^{-1}$ are due to acoustic oscillations in baryons. (Bottom) The differences between the power spectrum given by the BBKS approximation and the power spectrum obtained from our fits. Triangles show results obtained using COSMICS for $\Omega_b=0.05$
• Figure 2: Errors of the approximation eqs.(\ref{['EqUgly']}) for the CDM models with different Hubble constants and amount of baryons.
• Figure 3: The same as Figure 2, but for the CDM models.