- Scalable Generative Adversarial Networks for Multi-dimensional Images Ankit Srivastava, Nathanaël Perraudin, Aurelien Lucchi, Tomasz Kacprzak, Thomas Hofmann, Alexandre Refregier, Adam Amara

The dataset does not contain the Nbody simulations as they have a very large size. Instead, we sliced the space into 256 x 256 x 256 cubical areas and counted the number of particules in each area. The result are 3D histogram, where the number of particles is a proxy for matter density.

Note that a the same Nbody simulation were used in this paper, but with a different way of building the histogram.

- Fast Cosmic Web Simulations with Generative Adversarial Networks Andres C Rodriguez, Tomasz Kacprzak, Aurelien Lucchi, Adam Amara, Raphael Sgier, Janis Fluri, Thomas Hofmann, Alexandre Réfrégier https://arxiv.org/abs/1801.09070v1

N-body simulation evolves a cosmological matter distribution over time, starting from soon after the big bang. It represents matter density distribution as a finite set of massive particles, typically order of trillions. The positions of these particles are modified due to gravitational forces and expansion of the cosmological volume due to cosmic acceleration. N-body simulations use periodic boundary condition, where particles leaving the volume on one face enter it back from the opposite side.

We created N-body simulations of cosmic structures in boxes of size 100 Mpc and 500 Mpc with 512^3 and 1,024^3 particles respectively. We used L-PICOLA [21] to create 10 and 30 independent simulation boxes for both box sizes. The cosmological model used was ΛCDM (Cold Dark Matter) with Hubble constant H0 = 100, h = 70 km s−1 Mpc−1, dark energy density Omega_Lambda = 0.72 and matter density Omega_m = 0.28. We used the particle distribution at redshift z = 0.

$ cat run_parameters.dat

% =============================== % % This is the run parameters file % % =============================== %

% Simulation outputs % ================== OutputDir /cluster/scratch/jafluri/AndresBoxes2/Box_350Mpch_0/ % Directory for output. FileBase out % Base-filename of output files (appropriate additions are appended on at runtime) OutputRedshiftFile /cluster/scratch/jafluri/AndresBoxes2/Box_350Mpch_0/output_redshift.dat % The file containing the redshifts that we want snapshots for NumFilesWrittenInParallel 16 % limits the number of files that are written in parallel when outputting.

% Simulation Specifications % ========================= UseCOLA 1 % Whether or not to use the COLA method (1=true, 0=false). Buffer 3 % The amount of extra memory to reserve for particles moving between tasks during runtime. Nmesh 2048 % This is the size of the FFT grid used to compute the displacement field and gravitational forces. Nsample 1024 % This sets the total number of particles in the simulation, such that Ntot = Nsample^3. Box 350 % The Periodic box size of simulation. Init_Redshift 9.0 % The redshift to begin timestepping from (redshift = 9 works well for COLA) Seed 1020 % Seed for IC-generator SphereMode 0 % If "1" only modes with |k| < k_Nyquist are used to generate initial conditions (i.e. a sphere in k-space), % otherwise modes with |k_x|,|k_y|,|k_z| < k_Nyquist are used (i.e. a cube in k-space).

WhichSpectrum 2 % "0" - Use transfer function, not power spectrum % "1" - Use a tabulated power spectrum in the file 'FileWithInputSpectrum' % otherwise, Eisenstein and Hu (1998) parametrization is used % Non-Gaussian case requires "0" and that we use the transfer function

WhichTransfer 0 % "0" - Use power spectrum, not transfer function % "1" - Use a tabulated transfer function in the file 'FileWithInputTransfer' % otherwise, Eisenstein and Hu (1998) parameterization used % For Non-Gaussian models this is required (rather than the power spectrum)

FileWithInputSpectrum files/input_spectrum.dat % filename of tabulated input spectrum (if used) % expecting k and Pk

FileWithInputTransfer files/input_transfer.dat % filename of tabulated transfer function (if used) % expecting k and T (unnormalized)

% Cosmological Parameters % ======================= Omega 0.276 % Total matter density (CDM + Baryons at z=0). OmegaBaryon 0.045 % Baryon density (at z=0). OmegaLambda 0.724 % Dark Energy density (at z=0) HubbleParam 0.7 % Hubble parameter, 'little h' (only used for power spectrum parameterization). Sigma8 0.811 % Power spectrum normalization (power spectrum may already be normalized correctly). PrimordialIndex 0.961 % Used to tilt the power spectrum for non-tabulated power spectra (if != 1.0 and nongaussian, generic flag required)

% Timestepping Options % ==================== StepDist 0 % The timestep spacing (0 for linear in a, 1 for logarithmic in a) DeltaA 0 % The type of timestepping: "0" - Use modified COLA timestepping for Kick and Drift. Please choose a value for nLPT. % The type of timestepping: "1" - Use modified COLA timestepping for Kick and standard Quinn timestepping for Drift. Please choose a value for nLPT. % The type of timestepping: "2" - Use standard Quinn timestepping for Kick and Drift % The type of timestepping: "3" - Use non-integral timestepping for Kick and Drift nLPT -2.5 % The value of nLPT to use for modified COLA timestepping

% Units % ===== UnitLength_in_cm 3.085678e24 % defines length unit of output (in cm/h) UnitMass_in_g 1.989e43 % defines mass unit of output (in g/h) UnitVelocity_in_cm_per_s 1e5 % defines velocity unit of output (in cm/sec) InputSpectrum_UnitLength_in_cm 3.085678e24 % defines length unit of tabulated input spectrum in cm/h. % Note: This can be chosen different from UnitLength_in_cm