search

henoriega / FOLPS-nu

FOLPSν is a code for efficiently evaluating the redshift space power spectrum in the presence of massive neutrinos
8 stars 1 forks source link

option for output k-range #2

Open adematti opened 1 year ago

adematti commented 1 year ago

Hi, I am wrapping FOLPS into desilike. As an option, would it be possible to (manually) specify kTout to https://github.com/henoriega/FOLPS-nu/blob/f17f43cd08625c3ca072a0d4d5964ebb722c02dd/FOLPSnu.py#L643? For completeness I would like to provide the correlation function theory based on the power spectrum prediction, and the boundaries https://github.com/henoriega/FOLPS-nu/blob/f17f43cd08625c3ca072a0d4d5964ebb722c02dd/FOLPSnu.py#L673C36-L673C36 are a bit too tight. Of course, I could tweak things on my side, but this seems just very easy to add this option. Best, Arnaud

alejandroaviles commented 1 year ago

Hi, thanks.

I think is possible to enlarge the k range. However, when doing so the FFTLog method starts to break down. We can make some tests to check the maximum k for the results to still have sense. The other option is to exploit that at small scales all the functions basically follow a power law, so we can add a module to extrapolate these functions linearly in log-log space. But of course, this is will take some time to be done. What is the kmax we need to get a correlation function? Are you thinking of an FFTLog method?

adematti commented 1 year ago

I think best is just to add an option that allows the user to change the theory k or k-range (not change the default one). Tweaking kminout and kmaxout, the FOLPS power spectrum seems smooth up to 0.6 h/Mpc --- beyond this, we are extrapolating and damping with an exponential, and FFTlog to get a smooth correlation function. (The same routine is applied to other theory models).

alejandroaviles commented 1 year ago

Sounds good. @henoriega, could you add the an optional argument for the k-range? Of course, once you finish your exam in a couple of days

adematti commented 1 year ago

Hi @henoriega, is there any update on this? This is required to compare models in configuration space. I am happy to file a PR if that helps.