search

IvanNikolic21 / Lyman-alpha-bubbles

Codes used to derive the likelihood calculation for finding an ionized bubble given some Lyman-emitter data
0 stars 0 forks source link

Bins 8-10 bug #1

Closed IvanNikolic21 closed 4 months ago

IvanNikolic21 commented 7 months ago

There seems to be a problem with picking out the correct bubble configuration when using 8, 9 or 10 bins in the power spectrum. This happens even though bins 5,6,7 and then 11,12 and 13 work perfectly.

This is how the problem looks like with the green curve representing the problematic bin choice:

speed_up.pdf

There are several ways to tackle this problem, the best one being by looking at the corner plot to investigate what could be going wrong.

IvanNikolic21 commented 6 months ago

Note about this issue. There were some indications that this issue might not be present anymore. I will not close this issue for now, because I'm still waiting for the speed_up_bug_fix to finish before I can make the judgement.

IvanNikolic21 commented 6 months ago

Another update for this issue: speed_up_bug_fix has been merged and is done. I'm not going to close the issue despite the fact the bins seem to work. I want to check everything beforehand with the fully converged runs before.

IvanNikolic21 commented 4 months ago

I'm closing this issue as the bins that previously didn't work, outperform the ones that worked perfectly.