README
1$FreeBSD$
2
3Running:
41. Compile mergesort_bench.c into mergesort_bench
52. Run bench.py with python bench.py [elts]
62a. Bench will optionally run 2 ^ elts elements as the dataset size when provided. Otherwise it will run 2 ^ 20 elements.
7
8Output:
9Files will be output in a new folder called stats with separate files for each statistical comparison and the raw results in a subfolder called data.
10This files will contain the results of the running of ministat with time required to sort as the dataset.
11
12Modifications:
13Change bench.py's WIKI variable to be true if you have wiki.c implemented and want to test it.
14
15As the script runs, it is running each of the stdlib sorting algorithms (and wikisort if provided) with 2 ^ elts elements,
165 trials of the sort time as it's output. That output is saved in the data folder and then passed into the command line
17utility ministat which then provides the confidence interval of difference between the data in stats folder.
18