rFFT-based N1 lensing bias calculation and tests by Julien Carron. This code is tuned for TT, P-only or MV (GMV) like quadratic estimators.
-
rFFT-based N1 and N1 matrix calculations (~ in O(ms) time per lensing multipole for Planck-like config, which allows on-the-fly evaluation of the bias)
-
5 rFFT's of moderate size per L for N1 TT, 20 for PP, 45 for MV or GMV
This code is not particularly efficient for low lensing L's, since in this case one must use large boxes.
installation:
* pip install -e ./ [--user]
demo notebook here
Paper references: this code was developed for https://arxiv.org/abs/2206.07773 and https://arxiv.org/abs/2112.05764
- dN1/dcl TT, EE, BB tested for Planck and works below the percent for <= 400 and ~% for <= 2048, probably due to interpolation scheme
- case of l1 or l2 being zero still to fix