Computers and software • Re: Discrepancy between CAMB and CLASS for CPL parametrization

Hello,

I have studied the stability of derivatives on C_l to compute FIM elements with above formula.

I computed absolute derivatives as respect of Omega_m, w0 and wa (CPL parametrization).

You can see the following 3 figures :

1) The plot below of Omega_m is pretty good, there is stability for all
multipoles from 1e-4. I am a little surprised that the line for the multipole
l=20 is so low compared to that of l =100 ...



2) The plot of w_0 is also pretty good for h > 1e-4 but I have an instability that persists for the low multipoles (l=20 in blue) while
for higher multipoles, there is really good stability :



3) For the plot of w_a, then it is totally noisy : from 1e-8 to 1e-1, only oscillations :




I also attach in the archive the init file of CLASS (explanatory_base.ini).

I think I have a missing parameter into explanatory_base.ini or a fine-tuned value to modify.

Regards

Archive_Stability.tar

Statistics: Posted by Fabien Dournac — February 05 2025

---