See smmalis37/aoc2020@3fbf5d9#diff-6965df871359c91ee03b5564314f3e131c0877b8603aea626b5d8f9180eb4ee8R28-L29
The highlighted change, when measured in isolation from the other changes in that commit, resulted in a roughly 30-40% speedup of the entire function (i.e. the entire chain of iterators ending in the collect call that calls this FromIter implementation). If measured in isolation I assume the impact would be even greater. I expected the two to be equivalent.
The code being measured is day 11's parser: https://github.com/smmalis37/aoc2020/blob/3fbf5d9b9e89d18c213c0d9df6a08db9364b2604/src/days/day11.rs#L22. To replicate my results just clone that repo and run cargo run --release 11 and look at the measured time of the parser (the rest of day 11 will be measured too but it isn't calling this code).
See smmalis37/aoc2020@3fbf5d9#diff-6965df871359c91ee03b5564314f3e131c0877b8603aea626b5d8f9180eb4ee8R28-L29
The highlighted change, when measured in isolation from the other changes in that commit, resulted in a roughly 30-40% speedup of the entire function (i.e. the entire chain of iterators ending in the collect call that calls this FromIter implementation). If measured in isolation I assume the impact would be even greater. I expected the two to be equivalent.
The code being measured is day 11's parser: https://github.com/smmalis37/aoc2020/blob/3fbf5d9b9e89d18c213c0d9df6a08db9364b2604/src/days/day11.rs#L22. To replicate my results just clone that repo and run
cargo run --release 11and look at the measured time of the parser (the rest of day 11 will be measured too but it isn't calling this code).