lenscov

The package

This package contains scripts for computing CMB and lensing covariances as described in Peloton et al. 2016 (https://arxiv.org/abs/1611.01446) It has routines for:

Spectra
- computing (analytically) the N0 bias (curve- and flat-sky available)
- computing (analytically) the N1 bias (only flat-sky available)
- computing CMB lensed power spectra using series-expansion
- computing CMB lensed power spectra using correlation functions
Covariances
- computing (analytically) the covariance matrix of CMB lensed spectra up to second order in lensing potential power-spectrum
- computing (analytically) the covariance matrix of lensing potential spectrum
- computing (analytically) the cross-covariance matrix of CMB lensed spectra and lensing potential spectrum up to second order in lensing potential power-spectrum
Additional
- computing derivatives of spectra using correlation functions
- computing wigner (small) d-matrix efficiently

The code performs computation in the full-sky formalism, except the computation of N1 biases and their derivatives (see the LensingBiases package). The code is fully-parallelized using mpi4py, and some Fortran parts make use of openmp (computation of N1).

Before starting

This code has the following dependencies:

numpy, pylab, etc (required)
scipy (required, for C interfacing)
f2py (required, for Fortran interfacing)
mpi4py (optional, to use parallel computing)
LensingBiases (optional, to compute lensing biases)

Make sure you update your PYTHONPATH to use the code. Just add in your bashrc:

COVPATH=/path/to/the/package
export PYTHONPATH=$PYTHONPATH:$COVPATH:$COVPATH/src

Compilation of C and Fortran

The code is primarily written in Python, although the low-level routines (wigner3-j, nested loops, etc) are implemented in Fortran and C, and interfaced with Python. The compilation of C is done on-the-fly through scipy.weave. The compilation of Fortran is done either using the setup file, by running:

python setup.py build_ext --inplace --fcompiler=gfortran

Depending on system, you may need to specify a different fortran compiler. The compilation has been tested successfully with gfortran (gcc version 5.1.0) and ifort. For the purist, we also provide a Makefile for a direct compilation.

Usage

You can find a launcher (run_covariances.sh) to have a quick usage. We also provide a batch file for usage on cluster. You have pre-defined experimental configurations in misc.py.

License

GNU License (see the LICENSE file for details) covers all files in the lenscov repository unless stated otherwise.