This repository contains the main source code of the EuclidEmulator, a fast and accurate tool to estimate the non-linear correction to the matter power spectrum. EuclidEmulator is roughly seven orders of magnitude faster than an N-body simulation that yields results at the same level of accuracy.
Authors: M. Knabenhans & J. Stadel
Date of last update: September 2018
Reference: Knabenhans et al. (2019), MNRAS, 484, 5509-5529 (https://arxiv.org/abs/1809.04695)
If you use EuclidEmulator in any way (for a publication or otherwise), please cite this paper.
STAY TUNED:
- We are working on a new feature for the python wrapper allowing to easily and quickly calculate convergence power spectra.
- We plan a fully revised version of the emulator that includes neutrino physics.
If you have any questions and/or remarks related to this work, please do not hesitate to send me an email (mischakATphysik.uzh.ch)
In any case you need:
- GNU Scientific Library (GSL; see https://www.gnu.org/software/gsl/)
- CMake (see https://cmake.org)
In order to use the python wrapper, you need in addition:
- Python together with cython, numpy, scipy and pandas (only for python wrapper)
- CLASS code (see http://class-code.net) version 2.6.3 or higher together with the python wrapper classy (only for python wrapper)
If you have not done so already, either download this repository or clone it to your local host (under Linux you can get a gzipped tar-ball via
wget https://github.com/miknab/EuclidEmulator/archive/master.tar.gz
EuclidEmulator comes with a command line interface (CLI) and a full-fledged python package. You can choose whether you want to install only one of these tools or both. We recommend to install the python package as it's usage is less error-prone than that of the CLI and it offers much more utilities. In the following we will restrict our description on the python package (if you are interested in the CLI, please consult the wiki).
Now, inside the EuclidEmulator directory type
cd wrapper2if you want to install the python2 version of the EuclidEmulator wrapper, whereas if you want the python3 version you enter the wrapper3 directory like
cd wrapper3All the subsequent steps are identical for both python versions. Create a new build folder, enter it and run CMake, i.e.:
mkdir build
cd build
cmake ..(if you have root access) or
cmake -DCMAKE_INSTALL_PREFIX=path/to/installation/directory ..(without root access), where path/to/installation/directory denotes the absolute path to directory for which you have write access. Then type
make
make installin order to compile and install the code.
REMARK: If you don't have root acces on the machine you want to build this software on, we recommend to install this software inside a virtual environment (see documentation/wiki for more info).
Now you are ready to use the software. In a python script or in a jupyter notebook (with a kernel of the corresponding python version) you can import the package via
import e2pyNext, define a cosmology as a python dictionary with the keys om_b, om_m, n_s, h, w_0 and sigma_8:
MyCosmo = {'om_b': 0.0219961, 'om_m': 0.1431991, 'n_s': 0.96, 'h': 0.67, 'w_0': -1.0, 'sigma_8': 0.83}You can now compute a non-linear power spectrum at redshift z=0.0 by executing:
z = 0.0
result = e2py.get_pnonlin(MyCosmo, z)The result is also a python dictionary with the keys B (the boost factor, units: dimensionless), P_lin (the linear power spectrum as computed by class, units: (Mpc/h)3), P_nonlin (the non-linear power spectrum being the product of P_lin and BMpc/h)3) and k (the vector of k-values given in h/Mpc).
The result is also a python dictionary with the keys B (the boost factor, units: dimensionless), P_lin (the linear power spectrum as computed by class, units: (Mpc/h)3), P_nonlin (the non-linear power spectrum, also in units of (Mpc/h)3, being the product of P_lin and B) and k (the vector of k-values given in h/Mpc).
Notice that you can as well pass a list or numpy.ndarray for z in which case result will contain a nested dictionary: The fields B, P_lin and P_nonlin are now dictionaries themselves with fields z1, z2, z3, etc., corresponding to the boost factor, linear and non-linear power spectrum, respectively, evaluated at the individual redshifts listed in the z variable. Here is a short example:
import e2py
import numpy as np
MyCosmo = {'om_b': 0.0219961, 'om_m': 0.1431991, 'n_s': 0.96, 'h': 0.67, 'w_0': -1.0, 'sigma_8': 0.83}
z = np.linspace(0.0,1.0,11) # z = [0.0, 0.1, 0.2, ..., 0.9, 1.0]
result = e2py.get_pnonlin(MyCosmo, z)There will be e.g. the field result[B][z1] corresponding to the boost factor at redshift z=0.0 or the field result[P_nonlin][z9] corresponding to the non-linear matter power spectrum evaluated at redshift z=0.8.
NOTICE: The number of redshifts at which you ask EuclidEmulator to produce non-linear power spectra is limited. Usually this upper limit is 100 (sometimes already a bit less). This is a problem of CLASS (there is no such limit for boost factors whose computation does not rely on CLASS).
Credits for this project go to:
Doug Potter, University of Zurich, Switzerland for constantly advising me in this project; https://bitbucket.org/dpotter/
Jeppe Mosgaard Dakin, Aarhus University, Denmark for significant contributions to the cython code; https://github.com/jmd-dk
Rongchuan Zhao, University of Bonn, Germany for contributions in the lensing module ee_lens.py and some functions in ee_aux.py; https://astro.uni-bonn.de/m/rzhao.
EuclidEmulator is free software, distributed under the GNU General Public License. This implies that you may freely distribute and copy the software. You may also modify it as you wish, and distribute these modified versions. You must always indicate prominently any changes you made in the original code and leave the copyright notices, and the no-warranty notice intact. Please read the General Public License for more details.
- Python wrapper with many useful functions (reducing the risk of falling into a pitfalls)
- Usage of CMake for the building