This little project was made mostly because i wanted to catch up with sorting algorithms. Implementations are probably far from optimal so take it for what it's worth.
I tested 7 different algorithms:
1. Quicksort (not randomized)
2. Merge sort
3. Heapsort
4. Selection sort
5. Insertion sort
6. Shellsort (using Marcin Ciura's sequence)
7. Bubble sort
with 2 tests, first was taking an average of 100 sorts of randomized arrays with 5000 elements, second was mapping sorting time to size of an array to reveal differences in their complexity.
Important note: Both tests used time.clock() function which returns real time instead of cpu time on windows, so they're just an aproximation.
First test on my machine produced those results (rounded up to 4th decimal place):
1. Quicksort average = 0.0105 s
2. Heapsort average = 0.0141 s
3. Shellsort average = 0.0148 s
4. Merge sort average = 0.0216 s
5. Selection sort average = 0.6874 s
6. Insertion sort average = 1.522 s
7. Bubble sort average = 2.0728 s
Second test produced following graph:
I'm quite surprised how well Shellsort and how mediocre Merge sort did (since it should be O(nlogn) for every case) but my implementation may be suboptimal.