alexander-mead / bnl

Non-linear halo bias in the halo model

BNL

Example halo model code that includes non-linear halo bias following Mead & Verde (2020; https://arxiv.org/abs/2011.08858).

Clone the repository using

git clone --recursive https://github.com/alexander-mead/BNL

the --recursive is important because that will also ensure that necessary libraries are cloned in to the library/ subdirectory. BNL can then be compiled using >make. If you get an error: *** No rule to make target `build/precision.o', needed by `bin/BNL'. this is because you did not use the -- recursive flag above. BNL should compile with any Fortran compiler, the default is gfortran, but you can change the compiler within the Makefile. To run the compiled code type >./bin/BNL.

When run, the code should print some useful things to the screen and create data files in the results/ directory: power_std_linear.dat, power_std_2h.dat, power_std_1h.dat, power_std_hm.dat, power_bnl_linear.dat, power_bnl_2h.dat, power_bnl_1h.dat, and power_bnl_hm.dat. You should be able to plot these using the gnuplot script plotting/power.p, which is also included in the repository via >gnuplot plotting/power.p.

In the data/BNL/M512/ directory you will find measurements (in ascii format) of the BNL function from the Multidark simulation that are used in the code and paper. These measurements are in the *_bnl.dat files, which are two columns: k and 1+B_NL, k is in units of h/Mpc. The files are 1600 rows, which corresponds to 25 k points for each of the 8x8 mass bins (25 x 8 x 8 = 1600). The mass values corresponding to the 8 mass bins can be found in the *_binstats.dat files, which are 8 rows, one row per mass bin. The columns are: log10(m_min), log10(m_max), log10(m), nu_min, nu_max, nu, b and rv. Halo masses 'm' are in Msun/h and virial radius 'rv' is in Mpc/h, everything else is dimensionless, 'b' is the linear halo bias. The quantities without _min or _max denote an average over haloes in the bin, weighted by the number of haloes. The nu value is that associated with the mass bin in the code. In the *_bnl.dat files the mass bins are layed out sequentially like: 11, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, ..., 87, 88. There is some redundancy here since BNL is symmetric with respect to mass arguments, so for example 12 contains the same information as 21.

A simple Fortran loop to read a _bnl.dat file would be:

DO jbin = 1, nbin
  DO ibin = 1, nbin
     DO ik = 1, nk
        READ (u, *) k(ik), B(ik, ibin, jbin)
     END DO
  END DO
END DO

The different numbers for the *_binstats.dat and *_bnl.dat correspond to different Multidark snapshots. The redshifts are given in the MDR1_redshifts.csv file.

The research to develop this softare was funded by the Horizon 2020 research and innovation programme of the European Union under Marie Sklodowska-Curie grant agreement No. 702971.