Comparing Sort Algorithms

Reading that one sort is “n²” and another is “n log n” never did much for me. Watching them race does. So here they all are at once, running the same data, and you can see the fast ones lap the slow ones in real time.

Everyone takes the same number of comparisons per frame, so finishing order is really an operation count: the sort that does the least work crosses the line first. The clock next to each one is just along for the ride. There are two things worth watching: how fast it finishes, and how much extra memory it had to borrow to do it. The merge-style sorts are quick but they rent a whole second array to get there. The in-place ones are slower but barely touch a spare cell.

Pick which sorts line up, change the item count, choose random, sorted, or reversed data, and decide whether every sort gets the exact same array or its own shuffle. Then hit Play. New sorts drop in here the week they launch, so the field grows over time.