kbrooks87 / backend-benchmarking-activity

What time is it?

For this activity, you will be measuring the performance characteristics of various sorting algorithms, defined in sort.py.

Timeit

Experiment with the timeit module by seeing how long the various sort implementations take. While you could modify sort.py directly, you can take advantage of the command line interface to test different algorithms:

% python -m timeit "from sort import quick_sort as sort, random_seq; print(sort(random_seq))"

The reason for the as keyword is so that you can change search algorithms in one place instead of two.

After running the above command once, you can take advantage of your shell's command substitution by running the following in your terminal:

% ^quick^merge

The word after the first ^ signifies the word in the previous command to replace, and the word after the second ^ signifies the word to use as a replacement.

Doing so will result in the following command:

% python -m timeit "from sort import merge_sort as sort, random_seq; print(sort(random_seq))"

Special Note: When running your timing profile on the bubble_sort algorithm, don't give up! Your computer is not stalled in an infinite loop. Waaaaaiiiiiitttttt for it ....

cProfile

You can measure the performance of all algorithms by using the cProfile module and running the script:

% python -m cProfile -o sort.profile sort.py

You can then pass the resulting sort.profile file to the pstats module:

% python -m pstats sort.profile

Take a look at this article for help on how to analyze statistics.

Credits

For a closer look at some of the algorithms used in this activity, see A tour of the top 5 sorting algorithms with Python code.