(a python wrapper for ISiTGR version of CAMB has also been released -- see instructions further below)
We introduce a new version of Integrated Software in Testing General Relativity (ISiTGR) which is a patch to the software CAMB and CosmoMC. ISiTGR is intended to test deviations from GR at cosmological scales using available cosmological data sets. While doing so, it allows for various extensions to the standard flat LCDM model. In this new release, we have a combined support for the following:
-
Dynamical dark energy parametrizations with a constant or time-dependent equation of state;
-
A consistent implementation of anisotropic shear to model massive neutrinos throughout the full formalism;
-
Multiple commonly used parameterizations of modified growth (MG) parameters;
-
Functional, binned and hybrid time- and scale-dependencies for all MG parameters (expanded from previous version);
-
Spatially flat or curved backgrounds (present in previous version as well).
The description of the formalism and its implementation in the CMB code, the Integrated Sachs-Wolfe (ISW) effect, and the 3x2 point statistics as well as examples of application to current data sets, can be found in the latest paper on the arXiv. A more technical description of the implementation can be found in the documentation provided in this repository.
Additionally, this new version of ISiTGR is accompanied with a python wrapper. The python wrapper extends the original CAMB's python wrapper to work with the different MG parameterizations provided in ISiTGR, allowing the user to obtain the power spectra and transfer functions including phenomenological modified gravity parameters.
ISiTGR can be used as an extension of CosmoMC in order to perform MG and cosmological parameter inference or it can be used as a stand-alone extension of CAMB to produce cosmological observable considering the effect of MG parameters. In order to install ISiTGR the user will need a Fortran 2008 compiler such as Intel Fortran 14 or higher, or gfortran 6 at least. Find below the instructions about of how to install ISiTGR in each case.
The corresponding version of ISiTGR was built based on the CosmoMC July 2019. Installing ISiTGR as a patch of CosmoMC will allow you to perform MCMC sampling to constrain modified gravity and cosmological parameters. To install ISiTGR from the GitHub repository you can run the following commands in your terminal :
$ git clone https://github.com/mishakb/ISiTGR
$ cd ISiTGR
Select your compilers by editing the Makefile files inside the principal folder, camb/fortran folder and source folder. The original Makefiles are set to work in a particular set of clusters so you may need to modify them.
$ make
if there are no errors during compilation, you should see a new executable file called cosmomc and then you should be ready to use ISiTGR to run MCMC as described further below. You can verify that the installation was successful by running one of the tests files (test_ISiTGR.ini or test_ISiTGR_BIN.ini).
One can use ISiTGR as an extension to CAMB (version 1.1.0) without performing MCMC. ISiTGR as a patch to CAMB can be installed either by compiling the fortran code or downloading the new python wrapper. For the fortran base version of ISiTGR run the following commands in the principal folder:
$ cd camb/fortran
Select your compilers by editing the Makefile files inside the camb/fortran folder.
$ make
if there are no errors during compilation you should see a new executable file called camb, then you can test if the code is working by executing
$ ./camb ../inifiles/params_MG.ini
if no error is shown during the test then the code should be ready to go. Output is placed in the Spectra folder and more instructions about how to run ISiTGR-CAMB and parameters can be found in params_MG.ini file.
The ISiTGR patch allows to run CAMB and CosmoMC for MG different models. The next flowchart shows the different parameterizations and methods that ISiTGR is able to work with:
The user can take advantage of the ISiTGR capabilities after compiling the code by modifying the corresponding .ini files. The next flowchart shows the different .ini files that you need to modify in case you want to use the functional form or the binning methods for either CAMB (only use the files inside red circles) or CosmoMC (green circles for functional form and blue circles for binning methods). You can find further instructions inside each of this files.
It is important for the user to remember that the current version of ISiTGR is aimed to work with flat and non-flat models. Moreover, ISiTGR not only implements the contributions of massive neutrinos consistently, but also works with different equations of state for dark energy. The above mentioned is implemented for both functional form and binning methods.
If the you want to test the functional form of ISiTGR, you can to run
./cosmomc test_ISiTGR.ini
for a quick local run. Otherwise, if the you want to use binning methods, instead you should run
./cosmomc test_ISiTGR_BIN.ini
While ISiTGR-CAMB can be run locally to produce the spectra, using ISiTGR for MCMC sampling is recommended to be done on a cluster. We provide an example job script file inside the scripts folder that is designed to be run on a cluster with SLURM by just executing
sbatch scripts/run_ISiTGR.sh
Additionally, more instructions about the settings for the job can be found inside run_ISiTGR.sh
In the next table we show some plots to illustrate some features of the ISiTGR patch for CosmoMC:
Binning Methods | MG + Curvature | Planck 2018 data |
---|---|---|
If you want to use the Planck 2018 data, please follow the typical installation instructions as in CosmoMC.
In this new version, ISiTGR as a patch to CAMB can be run through a python wrapper in order to produce cosmological observables. The easiest way to install the ISiTGR python wrapper is using PyPI (https://pypi.org/project/isitgr/) by running
$ pip install isitgr [--user]
If you are working in a conda environment, then once you add conda-forge to your channels, ISiTGR is also available via
$ conda install isitgr
See also https://isitgr.readthedocs.io/en/latest/ for documentation and an example notebook for python-ISiTGR.
In the next table we show some plots to illustrate the different power spectrums that can be produced using the ISiTGR python wrapper:
Temperature and lensing potential power spectrum with (μ, η) | TT and TE power Spectrum with (μ, Σ) |
---|---|
Soon!
We would ask that when using ISiTGR or a modified version of it, you please cite our papers:
- ISiTGR: Testing deviations from GR at cosmological scales including dynamical dark energy, massive neutrinos, functional or binned parametrizations, and spatial curvature: https://arxiv.org/abs/1908.00290 (current ISiTGR version) by: Cristhian Garcia-Quintero, Mustapha Ishak, Logan Fox and Jason Dossett.
- Spatial curvature and cosmological tests of general relativity: https://arxiv.org/abs/1205.2422 by: Jason Dossett and Mustapha Ishak
- Testing General Relativity at Cosmological Scales: Implementation and Parameter Correlations: https://arxiv.org/abs/1109.4583 by: Jason Dossett, Mustapha Ishak and Jacob Moldenhauer
as well as the original CAMB paper; the original CosmoMC paper; Additionally please cite the use of any other datasets already included in the original version of CosmoMC.
Mustapha Ishak acknowledges that this material is based upon work supported in part by the U.S. Departmentof Energy, Office of Science, under Award Number DE-SC0019206 and the National Science Foundation under grant AST-1517768.
Cristhian Garcia-Quintero gratefully acknowledges a PhD scholarship from the Mexican National Council for Science andTechnology (CONACYT).
If you have comments, questions, or feedback, please feel free to contact to the contributors of this repository: [email protected] or [email protected].
Information from previous ISiTGR versions
We introduce a new version of Integrated Software in Testing General Relativity (ISiTGR) which is a patch to the software CAMB and CosmoMC. ISiTGR is intended to test deviations from GR at cosmological scales using available cosmological data sets. While doing so, it allows for various extensions to the standard flat LCDM model. In this new release, we have a combined support for the following:
-
Dynamical dark energy parametrizations with a constant or time-dependent equation of state;
-
A consistent implementation of anisotropic shear to model massive neutrinos throughout the full formalism;
-
Multiple commonly used parameterizations of modified growth (MG) parameters;
-
Functional, binned and hybrid time- and scale-dependencies for all MG parameters (expanded from previous version);
-
Spatially flat or curved backgrounds (present in previous version as well).
The description of the formalism and its implementation in the CMB code, the Integrated Sachs-Wolfe (ISW) effect, and the 3x2 point statistics as well as examples of application to current data sets, can be found in the latest paper on the arXiv. A more technical description of the implementation can be found in the documentation provided in this repository.
The corresponding version of ISiTGR was built based on the CosmoMC July 2018 version. To install the GitHub version of ISiTGR you can run the following steps in your terminal :
$ git clone https://github.com/mishakb/ISiTGR
$ cd ISiTGR
Select your compilers by editing the Makefile files inside the principal folder, camb folder and source folder.
$ make
$ make isitgr
if there are no errors during compilation, then you should be ready to use ISiTGR.
The ISiTGR patch allows to run CAMB and CosmoMC for MG different models. The next flowchart shows the different parameterizations and methods that ISiTGR is able to work with:
The user can take advantage of the ISiTGR capabilities after compiling the code by modifying the corresponding .ini files. The next flowchart shows the different .ini files that you need to modify in case you want to use the functional form or the binning methods for either CAMB or CosmoMC (you can find further instructions inside each of this files).
It is important for the user to remember that the current version of ISiTGR is aimed to work with flat and non-flat models. Moreover, ISiTGR not only implements the contributions of massive neutrinos consistently, but also works with different equations of state for dark energy. The above mentioned is implemented for both functional form and binning methods. If the user wants to run the functional form of ISiTGR, the user needs to run
./cosmo test_ISiTGR.ini
. Otherwise, if the user wants to use the binning methods, then one should run ./cosmo test_ISiTGR_BIN
.
In the next table we show some plots to illustrate some features of the ISiTGR patch for CosmoMC:
Binning Methods | MG + Curvature | MG + dark energy + neutrinos |
---|---|---|
In the next table we show some plots to illustrate the different power spectrums that can be produced using the ISiTGR patch for CAMB:
Angular power Spectrum for (μ, η) | Angular power Spectrum for (μ, η) allowing scale dependence |
---|---|
ISiTGR version used in https://arxiv.org/abs/1908.00290 to reproduce Planck 2015 results. The ISiTGR version 3.01 includes minor updates to work with the Planck 2018 data.
Integrated Software in Testing General Relativity
Version 2.01
Developed by Jason Dossett, Mustapha Ishak, and Jacob Moldenhauer.
What is ISiTGR?
ISiTGR is an integrated set of modified modules for the software package CosmoMC for use in testing whether observational data is consistent with general relativity on cosmological scales. This latest version of the code has been updated to allow for the consideration of non-flat universes. It incorporates modifications to the codes: CAMB, CosmoMC, and the ISW-galaxy cross correlation likelihood code of Ho et al. Also included is our independently developed generalized weak lensing likelihood module with data sets for the CFHTLenS weak lensing tomography of Heymans et al and CFHTLens 2D weak lensing measurements from Kilbinger et al.
A detailed explanation of the modifications made to these codes allowing one to test general relativity are described in our papers: arXiv:1109.4583, arXiv:1205.2422, and arXiv:1501.03119.
How to get ISiTGR
The two versions of ISiTGR have been consolidated into a single package. The three methods of evolving the parameters used to test general relativity, as described in our paper arXiv:1109.4583, are all contained within the code below. Different evolution methods are chosen by using different .ini files and changing options within those files.
This version of ISiTGR is for the July 2015 version of CosmoMC. The original (flat only) verison of ISiTGR as well as builds for other versions of CosmoMC are available here .
Running ISiTGR. (See further below ReadMe-II) To run ISiTGR you must first download and install CosmoMC (July 2015 version). You can find download instructions, system requirements, and setup instructions for CosmoMC in the CosmoMC ReadMe. Then, simply copy the contents of the unzipped ISiTGR folder to the CosmoMC folder, edit the Makefile for your compilers, compile using "make isitgr", and you are ready to go. To use the functional evolution of the modified gravity parameters, run the code with an ini file based on test_ISiTGR.ini. To use one of the binning methods run using an ini file based on test_ISiTGR_BIN.ini (Please Note: some changes to these ini files may be may be neccessary in order to run with the Planck 2013 likelihoods). Detailed explanations of the specific options available in ISiTGR are availalble in the ISiTGR ReadMe.
Referencing ISiTGR. We would ask that when using ISiTGR or a modified version of it, you cite: our papers (arXiv:1109.4583 and arXiv:1205.2422), this website; the original CAMB paper; the original CosmoMC paper; and the papers on original ISW-galaxy cross correlation likelihood by Ho et al and Hirata et al. When using any of the CFHTLenS weak lensing data sets, please cite our paper, the corresponding CFHTLens dataset paper (Heymans et al. or Kilbinger et al.), and follow the CFHTLenS publication guidlines given here. Additionally, please cite the use of any other datasets already included in the original version of CosmoMC.
Version History 2.01 (Released 07/23/15, this version) Updated to July 2015 version of CosmoMC. 2.0 (Released 05/07/15), Major release upgrade: Consolidated the two versions of the code into a single package, major updates to the likelihood modules for compatibility with the Feb. 2015 version of CosmoMC. 1.2 (Released 01/25/15): Updates for Dec. 2013 version of CosmoMC, new CFHTLenS likelihood module. 1.1.1 (Released 05/29/13): Includes bug fixes (Special thanks to Ana Caramete and Lucia Popa). 1.1: Updated code to allow for non-flat universes when varying the MG parameters. 1.02.1 (Released 04/27/12): Fixed bug in CMB_Cls_simple.f90. 1.02: Updated to January 2012 version of CosmoMC. 1.01.1 (Released 04/27/12): Fixed bug in CMB_Cls_simple.f90. 1.01: Updated to October 2011 version of CosmoMC. BAO module now comes included in CosmoMC (Thanks to Antony Lewis). 1.0: Initial release, for August 2011 version of CosmoMC. Contact If you have any question or comments about ISiTGR or would like to be updated about with future changes to the code, please feel free to email Mustapha Ishak at: [email protected] and Jason Dossett at: [email protected].
Readme-II : further info
Introduction
ISiTGR is an integrated set of modified modules for the software package CosmoMC for use in testing whether observational data is consistent with general relativity on cosmological scales. This latest version of the code has been updated to allow for the consideration of non-flat universes. It incorporates modifications to the codes: CAMB, CosmoMC, and the ISW-galaxy cross correlation likelihood code of Ho et al. Also included is our independently developed generalized weak lensing likelihood module with data sets for the CFHTLenS weak lensing tomography of Heymans et al and CFHTLens 2D weak lensing measurements from Kilbinger et al.
To use ISiTGR you must first download and install CosmoMC (July 2015 version). You can find download instructions, system requirements, and setup instructions for CosmoMC in the CosmoMC ReadMe. Then, simply copy the contents of the unzipped ISiTGR folder to the CosmoMC folder, edit the Makefile for your compilers. Compile using the command make isitgr, and you are ready to go.
Basic usage of ISiTGR is the same as CosmoMC. See below for extra options in the .ini files.
Differences between ISiTGR and CosmoMC Other than the changes to the code to incorporate the parameters used to test general relativy as described in our papers: arXiv:1109.4583 and arXiv:1205.2422, the following additional changes have been made
ini files There are two input ini files provided with ISiTGR. test_ISiTGR.ini allows you to run the code with the MG parameters evolved using our the functional form evolution. test_ISiTGR_BIN.ini allows you to run the code with the MG parameters evolved using one of the binning methods. The following options are available in boths of these ini files. To run the ISW-galaxy cross correlation likelihood code, uncomment the line DEFAULT(batch2/ISWHo.ini) in the test_*.ini file that you are using.
To use the CFHTLenS weak lensing tomography data set simply make sure the line DEFAULT(batch2/CFHTLens.ini) is uncommented in test.ini. This file enables use of the full CFHTLenS data sample as described in Heymans et al. If you would like to use one of the other galaxy samples instead you can change that line to: "Early Type," Red sample - DEFAULT(batch2/CFHTLens_red.ini) "Late Type," Blue sample - DEFAULT(batch2/CFHTLens_blu.ini) "Optomized Early Type" sample - DEFAULT(batch2/CFHTLens_rfbb.ini) These files each contain various settings to enable the different data sets as well as the default parameter ranges for the CFHTLenS intrinsic alignment (IA) nuissance parameter, ACFHTLenS, which allows this parameter to vary from -10 to 10. If you would like to fix this parameter to 0 and thus ignore the (IA) signal you can add the following line at the end of test.ini: param[CFHTLensA] = 0
To use the CFHTLens 2D weak lensing data set from Kilbinger et al, simply add the line DEFAULT(batch2/CFHTLens_2D.ini) to your test_*.ini. Additional parameters and options for testing general relativity: Functional Evolution The file batch2/common_ISiTGR.ini contains lines the line: parameterization = ISiTGR which tells CosmoMC that we want to use the functional evolution of the MG parameters.
The following options have been included in batch1/params_ISiTGR_defaults.ini
The option: Use_R_Function = lets you decide which parameters for testing GR you want to evolve using the funcitonal form shown in our paper. Setting this option to T evolves the parameters Q and R, while setting this option to F evolves Q and Σ (previously called D). For varying the parameters used to test GR we add the following parameter lines, please refer to our paper for a detailed explanation of each of these parameters. param[Q0] = 1 0 10 0.05 0.05
param[Qinf] 1
param[Sigma0] = 1 0 10 0.05 0.05
param[Sigmainf] = 1
param[kc] = 0.01
param[s] = 0 0 3 -1 1
The option: Scale_Dependent = lets you decide whether the evolution of the parmaters is scale dependent. Setting this option to T enforces scale dependence, and the parameters ending in inf above should be varied or changed to whatever value you want the parameters to take on small scales.
Additional parameters and options for testing general relativity: Binning Methods The file batch2/common_ISiTGR_BIN.ini contains lines the line: parameterization = ISiTGR_BIN which tells CosmoMC that we want to use the binning methods to evolve the MG parameters.
The following options have been included in batch2/params_ISiTGR_BIN_defaults.ini
For varying the parameters used to test GR we add the following parameter lines, please refer to our paper for a detailed explanation of each of these parameters. param[Q1] = 1 0 10 0.05 0.05
param[Q2] = 1 0 10 0.05 0.05
param[Q3] = 1 0 10 0.05 0.05
param[Q4] = 1 0 10 0.05 0.05
param[Sigma1] = 1 0 10 0.05 0.05
param[Sigma2] = 1 0 10 0.05 0.05
param[Sigma3] = 1 0 10 0.05 0.05
param[Sigma4] = 1 0 10 0.05 0.05
param[kc] = 0.01
The option: z_grid_spacing = defines the size of your bins in redshift bins. The option: Do_Exponential_Binning = lets you choose how you want to transition between scale bins. Setting this option to T sets transitions between scale bins to behave as an exponential function with a decay constant kc (This is the hybrid method described in our paper). Setting this option to F sets transitions between scale bins to behave as a hyberbolic tangent functions (near step functions) with the bins divided at k = kc.
ISW-galaxy cross correlation likelihood code To implement the ISW-galaxy cross correlations likelihood code of Ho et al, the following files have been added to source/: lrg_2dCl.f90 and iswdata.f90 . The files jl.f90 and lrg_pk.f90 have been removed in this latest version of the code. The folder bdndz_code which contains c-codes called by the routine Iswlnlike in iswdata.f90 has been added to the root directory. These c-codes are compiled by running the make isitgr command from the top level cosmomc directory. When Iswlnlike calls the c-program dndz.x it passes input files and output file names to the program. By default, these files are saved in the IO directory within the bdndz_code directory, however you can enable writing of these files to a different directory by setting the environmental variable TMPDIR.
Generalized Weak Lensing Likelihood module and Codes ISiTGR includes a generalized weak lensing likelihood module, source/WeakLen_Common.f90. This module defines the derived type TCosmologyWLLikelihood and various associated procedures. From this derived type, one can define specific routines for various weak lensing likelihoods. The specific routines relating to the CFHTLens data sets are contained in CFHTLens.f90. All of these files are located in the source directory. These files are loosely based upon the files contained in the original COSMOS weak lensing likelihood code of Lesgourges et al.
Referencing ISiTGR We would ask that when using ISiTGR or a modified version of it, you cite: our papers (arXiv:1109.4583 and arXiv:1205.2422), this website; the original CAMB paper; the original CosmoMC paper; and the papers on original ISW-galaxy cross correlation likelihood by Ho et al and Hirata et al. When using any of the CFHTLenS weak lensing data sets, please cite our paper, the corresponding CFHTLens dataset paper (Heymans et al. or Kilbinger et al.), and follow the CFHTLenS publication guidlines given here. Additionally please cite the use of any other datasets already included in the original version of CosmoMC.