lbj-wade / n-genic_growth Goto Github PK

This is the N-GenIC code used to generate the N-body initial conditions.
The codes uses the Zel'dovich approximation to generate the IC.
The code can create the ICs for pure CDM simulations, simulations containing
CDM & Baryons, CDM & neutrinos and CDM Baryons & neutrinos.

To compile the code the flags

-DMULTICOMPONENTGLASSFILE
-DDIFFERENT_TRANSFER_FUNC
-DNEUTRINOS

have to be switched on, even for pure CDM simulations.

For simulations with no baryons, set OmegaBaryon to 0 in the N-GenIC parameter 
file.

We have modified the files:

-allvars.h
-allvars.c
-read_param.c
-main.c

to allow the user whether or not Rayleigh sample the modes amplitude.


The code is able to deal with k-dependent growth factors, as those present in 
massive neutrino cosmologies. In the parameterfile the file containing the 
growth = dlog(delta) / dlog(a) is set with the parameter FileWithGrowth. That
file should be computed previously from two close power spectra from CAMB.

The value of H(z_initial) should also be input in the parameter file under 
the parameter.

The value of sigma8 in the parameterfile is just a sanity check if the
parameter ReNormalizeInputSpectrum is set to 0.

The code has the functionality to output the linear density field used to
generate the initial conditions. For this switch on the OUTPUT_DF flag in
the Makefile. The output will be three different files

-Coordinates_ptype_X.Y
-Amplitudes_ptype_X.Y
-Phases_ptype_X.Y

where X can be 0 (gas),1 (cdm) and 2 (neutrinos), and Y goes from 0 to the
effective number of cpus used when running the code

All the files have a header:
Nfiles(int32) Nmesh(int32) Local_nx(int32)

The header is followed by the data from Local_nx*Nmesh*(Nmesh/2 + 1) modes
In Coordinates is np.int64 and in Amplitudes and Phases is np.float32

Coordinates is a single long integer containing the position of the mode:
Coordinate = (i*Nmesh + j)*(Nmesh/2 + 1) + k, where i and j go from 0 to Nmesh
and k goes from 0 to Nmesh/2 + 1

Amplitudes contains a float with the value of the amplitude (A from A*exp(i*theta))
The units can be obtained by multiplying its number by BoxSize**(3.0/2.0)

Phases contains a float with the value of the phase (no units)

The Pk_library.pyx of the Pylians library contain a routine (Pk_NGenIC_IC_field)
to compute the Pk from these files