cola_halo

Parallel COLA cosmological simulation + 2LPT initial condition generator + FoF halo finder

This is a parallel cosmological N-body simulation code running the following on the fly, designed to generate hundreds of realizations:

1. Generates random Gaussian initial condition with 2LPT -- second-order Lagrangian Perturbation Theory (Scoccimarro 1998; Crocce, Pueblas, & Scoccimarro 2006,[1]).
2. Time evolve N-body particles with the COmoving Lagrangian Acceleration (COLA) method (Tassev, Zaldarriaga, & Eisenstein 2013,[2])
3. Finds dark-matter haloes with the Friends-of-Friends algorithm (Davis et al. 1985, Washington University N-body shop).

This code is based on publicly available codes (1, 2, 3), modified and redistributed under GPLv3.

Compile

This code requires an MPI C compiler mpicc, GNU Scientific Library GSL and FFTW. If they are installed in nonstandard locations, set the variables GSL_DIR and FFTW3_DIR in the Makefile.

% make

should compile the code and create the executable cola_halo.

Run

• The simulation requires a linear power spectrum extrapolated to redshift z=0, specified by the powerspectrum variable in the parameter file param.lua. The file should give a list of k [h/Mpc] and P(k) [(Mpc/h)³], which can be the matterpower.dat file generate by CAMB. The power spectrum should have the σ₈; specified in the parameter file (the code does not rescale the amplitude to the given σ₈). A power spectrum file consistent with the default parameter file is here.

• Edit the parameter file param.lua and set the number of particles, box size, number of time steps, etc.

• Run the code with the name of the parameter file

   mpirun -n 4 cola_halo param.lua

Where 4 is the number of MPI nodes you use. It can be any number, but we recommend it is a divisor of the number of particle per dimension nc for effective domain decomposition.

Outputs

File names are formatted as fof00100a.txt, where 00100 is the random_seed for the initial condition, and a,b,c,... corresponds to first, second, third, ..., output_redshifts.

• The N-body particles snp are in the GADGET binary format.
• The halo catalogue fof is a ascii file:

nfof x y z vx vy vz

where nfof is the number of N-body particles in the FoF group, (x y z) is the centre-of-mass position in comoving 1/h Mpc, (vx vy vz) is the average peculiar (physical) velocity in km/s.

• The matter density field grid and subsampled particles sub are simple binary files; see tools/density_grid.c and tools/subsampled_particles.c.

Options

• You can do simple calculation in the parameter file, e.g,

coarse_grid_nc = nc/4

with the LUA programming launguage by turning on the USE_LUA option in the parameter file.

• You can turn on Hybrid MPI+OpenMP parallelisation by setting the OPENMP variable. The FoF halo finder, however, is not parallelised by OpenMP.

Related codes

• l-picola: Another parallel COLA implementation by Cullan Howlett and Marc Manera 4, including on-the-fly lightcone generation.

• QrPM: A fork of this code. The parallelisation improved by using 2D domain decomposition. 5

Authors

This code is assembled and parallelized (COLA & FOF) by Jun Koda. See reference 5 below about the application to the mock generation of WiggleZ Dark Energy Survey and its accuracy. Feel free to open an issue for bug reports or questions.

References

1. Crocce, M., Pueblas, S., & Scoccimarro, R. 2006, MNRAS, 373, 369
2. Davis, M., Efstathiou, G., Frenk, C.~S., & White, S.D.M. 1985, ApJ, 292, 371
3. Scoccimarro, R. 1998, MNRAS, 299, 1097
4. Tassev, S., Zaldarriaga, M., & Eisenstein, D.~J. 2013, JCAP, 6, 36
5. Koda, J., et al 2015, arXiv:1507.05329, MNRAS submitted