NHSE Software Catalog

Cosmological Simulations Using Adaptive Particle-Mesh Methods

url

ftp://ftp-hpcc.astro.washington.edu/pub/hpcc/adap.sh.Z

author

Hugh Couchman, University of Western Ontario

contact

couchman@cita.utoronto.ca

abstract

The aim of cosmological particle codes is to follow
the evolution of a large number of interacting particles under
Newtonian gravity.
Calculating interparticle forces with a regular mesh results in an
equivalent minimum particle softening of roughly two mesh spacing.
this is the basic Particle-Mesh (PM) calculation. It is suitable
for the study of many collisionless phenomena, such as the dynamics
of galaxies and, in the context of cosmology, for the study of
the topology of large-scale structure. Modelling of collisional
systems, such as the clustering and dynamics of point-mass galaxies,
requires a softening significantly smaller than the mean interparticle
spacing. It is generally uneconomical to provide the required force
by resorting to increasingly fine meshes.
The Particle-Particle-Particle-Mesh ($P^3M$) method of Hockney
and Eastwood provides a feasible alternative. This method augments
the mesh force on a particle with a short range component summed
directly from near neighbours.
The particle-particle (PP) calculation must search for neighbours
out to roughly two mesh spacings to properly augment the PM force.
The number of particles within this search distance grows as
clustering develops, significantly slowing calculation of the PP sum.
the simples extension to $P^3M$ is to use a finer mesh in regions
where the particle density is high, in order to reduce
the number of particles per mesh cell back to unity.
The Adaptive $P^3M$ technique ($AP^3M$) is built around the standard
$P^3M$ algorithm, which itself consists of the PM cycle followed by
the direct particle-particle sum. $AP^3M$ produces fully equivalent
forces to $P^3M$ but represents a more efficient implementation
of the force spltting idea of $P^3M$.

description

http://www-hpcc.astro.washington.edu/simulations/DARK_MATTER/adapintro.html

reference

http://www-hpcc.astro.washington.edu/simulations/DARK_MATTER/adap.ps

keywords

cosmology; particle mesh method; application program