Advent_of_code_2021 VS aoc2021

holy shit, your solution is amazing. It's so clear. It took me a day and a really complicated solution to get the answer in more than 5 mins runtime (https://github.com/Harshsa28/Advent_of_code_2021/blob/main/src/22.py).

With only a few small restrictions down from "fully general", I have a brute-force-ish solution in python running in about 15 seconds on my hardware and my input: https://github.com/fizbin/aoc2021/blob/main/aoc24b.py

With day 22, I actually have three different python approaches in my repo; the inclusion/exclusion approach is what I named aoc22c.py.

This sounds vaguely similar to the approach I took on my second python solution, minus one optimization that I used: namely, I determined the maximum allowable value of "z" at the end of each instruction if we were going to end the program with z=0 and used that to throw away states that weren't worth pursuing.

I've implemented that approach, and my code to do so takes less than five seconds on my machine, so I wonder: have you tried this on the sample input? What does it do there?

How about a python translation of my haskell solution?

My solution in my github repo

Here's the more straightforward Monad-based solution, which unfortunately takes well over a minute to run.