#
ECCA Software

The
gravitational equations are evolved for a collisionless gas
using a Particle-Mesh solver on a Cartesian grid (Meiksin
& White 2001). The code is coupled to a radiative transfer
code *PMRT* using a probabilistic method to represent
absorption of photons
based on the photon-conserving algorithm of Abel, Norman & Madau
(1999, ApJ, 523, 66), extended to include helium by Bolton,
Meiksin & White (2004). The code is described in
Tittley &
Meiksin
(2006).Both the gravitational and radiative modules are coded in C
and
parallelised using a message passing interface (MPI).
Recently we have merged the radiative transfer algorithm with the
Adaptive Mesh Refinement numerical hydrodynamics code
*Enzo* (Bryan
& Norman 1997; O'Shea, Bryan,
Bordner, Norman, Abel, Harkness, & Kritsuk 2004), and are
currently running tests of the combined code *EnzoRT*.

A Java GUI that computes the amount of intergalactic extinction due to scattering and absorption
by neutral hydrogen in the IGM and its effect on galaxy colours is available here. A tarball for the code may be downloaded from tarball.

The
code used is a combined gravity-hydrodynamics scheme. The
equations of hydrodynamics are evolved using Smoothed Particle
Hydrodynamics (SPH). The gravitational equations are evolved using a
tree structure to compute the gravitational forces and the nearest
neighbours of particles. Particles are advanced with individual
timesteps, resulting in a huge computational savings for problems
involving a wide range of dynamical time scales. The code is in
Fortran 77 and parallelised using OpenMP. The code is described in
Bate,
Bonnell & Price (1995).

The equations of hydrodynamics are evolved using a shock-capturing,
explicit, finite difference algorithm, the Piecewise Parabolic Method
(PPM). The Newtonian potential of all matter on the grid is computed
using Fast-Fourier Transform (FFT) routines. Gravitational wave
emission and back-reaction are implemented in a post-Newtonian
approximation. The equations are solved on multiply nested and
refined Cartesian grids designed to focus on the regions requiring
short time scale and high spatial resolution. Both the Lattimer &
Swesty and the Shen et al. equations of state are included, as is
neutrino emission. Neutrino-antineutrino annihilation is handled in a
post-processing stage. The code is written in Fortran 77 and
parallelised using OpenMP. A short description of the numerical
procedures, initial conditions and results can be found in an extended
abstract (gzip'd
postscript, 37KB), which is a contribution to the
Proceedings of the 17th Texas Symposium (Munich, Dec.11-17 1994).

The work in this area uses standard N-body algorithms (the Aarseth
code
*N-Body*
and the *Starlab*
package) and also an adaptation of the Warsaw Monte Carlo code of Mirek
Giersz
(
2006, MNRAS, 371, 484: "Monte Carlo simulations of star
clusters - III. A million-body star cluster").