News for hackers. Stuff that matters.
Re:Speeding up my Advent of Code solution with Dijkstra's Al... by yxhuvud on Saturday January 01, 2022 @08:16AM (#29758026)
Yes, though sometimes it is possible to replace a fully general priority queue with a faster structure that is using the structure of the problem at hand. For example, this problem (AoC day 15) has only ever a count of unique priorities that is 11 or fewer. That allows a queue implementation to be https://github.com/yxhuvud/aoc21/blob/main/day15.cr#L27-L51 , being amortized O(1) in both insertion and deletion. This pushed down the runtime another magnitude for me, the whole day running at 0.014s.

There are no comments on this comment yet.