Tips
Speed
Spaceborne offers the possibility to cache the result of the most time-consuming
operation – the computation of the cNG connected trispectrum
\(T^{ABCD}(k_1, k_2, a)\) – for later use. When running the code for the first time, set:
PyCCL:
load_cached_tkka: False
To avoid running into errors (since the file does not exist yet). If you rerun the code
with consistent settings, you can load this in later runs by simply changing the
above flag to True.
Warning
The code will not check the consistency of the configuration used to generate the cached files when loading them for subsequent runs. If, for example, run 2 loads the cached trispectrum from round 1 but has a different halo model, the code will produce inconsistent results.
Note: these and other expensive operations are run in parallel, so the code will run faster simply by increasing the number of threads used:
misc:
num_threads: 40