Lensed CMB v0.5 improvements

ACTCollaboration / actsims

Simulations of the microwave sky as seen by ACT and Planck.

9 stars 4 forks source link

Closed msyriac closed 3 years ago

msyriac commented 3 years ago

This PR introduces several improvements to the lensed CMB generation:

higher accuracy settings, adequate given foreground levels (see plots below); the increase in CPU time is not substantial
logging to a file with more git info for both actsims and pixell
command line arguments with defaults read from a config file

ccl rcl

Note that since the filename convention has already changed in previous PRs, I went ahead and made the spelling corrections we wanted.

msyriac commented 3 years ago

The vertical dashed line marks ell=8250, beyond which the lensed CMB prediction in our CAMB files is zero.

ajvanengelen commented 3 years ago

Thanks Mat! This all looks great! A few (very minor) thoughts

I am sad to sad my cute little usefulInfo routine go, but I see you replaced it with something much more professional-looking
- Not sure you saved all the same info (total runtime, git commit hash)
Also, more scientifically, what did you do to increase (not decrease) the total power?
And are we OK with the 0.1% excess at l > 8100 (my guess is yes we are, but it is a question for the PS group)
Finally, I didn’t know about the actsims.util to get the rank and sizefrom mpi4py, thanks

ajvanengelen commented 3 years ago

Thanks for answering these questions offline. Merging