__ __ __ ____ __ _____ __ ________
/ // /_ _____ ___ ____ / |/ (_) /_____ ___ / ___/_ __/ / ___ /⠡ /\
/ _ / // / _ \/ -_) __/ / /|_/ / / '_/ -_|_-< / /__/ // / _ \/ -_) / ⠡..../..\
/_//_/\_, / .__/\__/_/ /_/ /_/_/_/\_\\__/___/ \___/\_,_/_.__/\__/ /___⠌___/ /
/___/_/ \ ⠌ \ /
An experimenal project \⠌______\/
https://www.youtube.com/watch?v=g9n0a0644B4
Given a number of cubes, how many ways could you uniquely arrange them?
- After generating an arrangment of cubes, we will compute it's hash.
- The number of consequative zeros at the start of the hash is the score.
- The last 8 bits mod num_buckets is the bucket to choose from.
- If the score is larger than the selected bucket's score, replace it.
- Repeat...
- Profit???
- Estimated arrangements = 2 ^ harmonic mean of buckets.
This results in a probabilistic method (which is suboptimal) of finding the number of cubes, but comes with far less coding, complexity, easy multithreading/scaling, and imo is "good enough" to at least get an estimate for larger numbers of cubes.
- Rotation and flipping is not accounted for.
- It still takes quite a bit of computation.
- I have a potato as a PC.
- Have no idea when it converges (if it does?)
- Zero regard for multithreading, etc, just an experiment.
todo