sigma8 consistency test

[YaoQianli]

check emulator's consistency with CAMB
%time sigma8 = csstemu.get_sigma8(type='CAMB')
print('CAMB, sigma8 =',sigma8)
%time sigma8 = csstemu.get_sigma8_cb(type='Emulator')
print('emulator no neutrino, sigma8 =',sigma8)
gives the output:

Different methods have a 0.003 difference in sigma8.
Possibly integration accuracy?

CSST WL alone can constrain sigma8 at ~0.01, considering we normally require 0.3sigma accuracy, a 0.003 difference in sigma8 reaches the limit. So it should be identified before combining CSST WL data and other observations.

[czymh]

This inconsistency is caused by the comparison between sigma8 and sigma8_cb.
These two functions return the same result only when the mnu=0, which means there is no massive neutrino and matter only includes cold dark matter and baryon (indicated by 'cb').
If mnu > 0, you should always use the csstemu.get_sigma8 to obtain the sigma8.
I have updated the package and added the neutrino linear power spectrum emulation to support the sigma8 calculation for the type of "Emulator" with a much shorter time (**~ 5ms**).
So that you can compare the 'Emulator' type with 'CAMB' or 'CLASS'.
The difference between 'Emulator' and 'CAMB' or 'CLASS' is less than 0.1%.
The results are shown in the following figure, detailed in test/test-sigma8.ipynb.