jk-jeon / floff Goto Github PK

I'm trying to recreate lookup tables for the algorithm, and it seems like the first thing to do is to calculate Q for all of the $(e, k)$ pairs. I wrote the following script to do so. The problem is that it returns smaller values for $Q$ than the ones mentioned in your blog post. For example, it returns $Q=110$ for $\eta=1$ and, what is more interesting, $Q = 183$ for $\eta = 23$.

get_bounds function, which is responsible for finding best rational approximations, seems to pass stress testing (see lines 77-90), so I'm fairly certain that it works. However, I'm clueless how to debug this further, since I haven't found code which calculates original values in the repo.

I know that this issue looks like "please, debug my random piece of code", so feel free not to help me and close the issue :)