Open j-murphy-slingshot opened 1 year ago
Isn't that what peakfinder
is for?
I don't think so, to me it looks like
peakfinder
is a function that can find peaks within a bounded subregion of the integrand
whereas the phase argument allow for a different integrand during the phase 0: sampling and phase 1: partitioning of the Divonne algorithm to assist with efficiency & potentially accuracy during the tesselation.
To me it looks like they can be in conjunction to generate a more robust and efficient integration, but am open if you have a different interpretation.
On page 14 of the Cuba paper (https://arxiv.org/abs/hep-ph/0404043) there is discussion about having a phase argument on the integrand so that an approximation with similar peak structure can be passed for the beginning of the algorithm. There doesn't seem to be support on the Julia interface for accessing this functionality