Simple and unnecessary LogVal class for C++.
The LogVal class uses logarithms to express large numbers and do math with them. Basically it is a wrapper around these identities:
This leads to major mathematical inaccuracies but makes it possible to run
algorithms where fixed-width floating point numbers (double, float) are
too small and arbitrary precision floating point numbers are too slow.
Initially, I used that to implement a FFT method to estimate the density of states of physical systems. Honestly, I found no other use-case for this class, but reimplementing it gave me the chance to try out C++20 stuff.