There was a series of number ranges that mapped onto other ranges. The simple approach was filling dictionaries, which worked well for the example data. In the actual data, there ranges were much much larger (in the 100,000,000’s), which made the dictionaries prohibitively large.
![[2023 Day 6] No optimization needed](https://programming.dev/pictrs/image/defe3416-b17d-40e9-a786-94f0b42a7e9e.png){kind=link}
There was a series of number ranges that mapped onto other ranges. The simple approach was filling dictionaries, which worked well for the example data. In the actual data, there ranges were much much larger (in the 100,000,000’s), which made the dictionaries prohibitively large.
If each number was 8 bytes wide, it would require, if I didn’t mess up my math, 18 GB to fully represent all seeds/numbers as an array.