litebird / litebirdxs4 Goto Github PK

Currently maps.get_s4_map fetches CMB-S4 LAT CMB-only maps at NSIDE=4096. The noise maps are at NSIDE=2048, while the lensed CMB is computed till lmax=4096. So we do not need to fetch the CMB-only LAT maps at NSIDE=4096 and increase compute time for each iteration to produce coadded maps. Can we switch to a default of NSIDE=2048 for the LATs?

Currently the simulations are split in at least two different format (fits, npy). We require two different python packages maps and chwide_maps to query the npy files at runtime. Unforturnately the information about these packages and their dependencies are unclear and scattered between multiple sections. It would be helpful if there is a master document clarifying what are the minimal set of steps and dependecies that are needed to access the simulations. Note that maps package seems to require two instrument model tables which one has to trace down. I still don't have the Litebird table. Can all of these be put together in a single package with a clear documentation.

The noise and foreground for CMB-S4 are in Equatorial coordinates while for LiteBIRD are in Galactic coordinates. The CMB maps fetched through maps package needs a relative rotation between the LiteBIRD and CMB-S4 CMB maps. This is cheaper to implement on the LiteBIRD maps as they are at a lower resolution (hence lower lmax). We need a rotate_alm stage in get_lb_map to convert from Equatorial to Galactic coordinates.

@antolonappan Please add any requirements that I may have missed.

@nraffuzz @toshiyan

please make the files in

/global/cfs/cdirs/cmbs4xlb/v1/noise/lb/ owned by the group cmbs4xlb

also the noise matrices are unaccessible

@giuspugl do you have a standard Nside you run your simulations at? either same or different for each channel.
For comparison, S4 does 512 for SAT 4096 for LAT, SO varies by channel: https://github.com/simonsobs/map_based_simulations/blob/master/mbs-s0012-20230321/simonsobs_instrument_parameters_2023.03/so_default_resolution.csv