profiling_code

Profiling Code

• way to measure the performance of your code
• params like: execution time, memory usage, CPU usage, etc.
• identify bottlenecks and optimization opportunities

cProfile

• built-in Python module for profiling code
• usage: python -m cProfile main.py
• expect results like:

7 function calls in 45.005 seconds

Ordered by: cumulative time

ncalls  tottime  percall  cumtime  percall filename:lineno(function)
    1    0.000    0.000   45.005   45.005 {built-in method builtins.exec}
    1    0.000    0.000   45.005   45.005 main.py:1(<module>)
    1   45.005   45.005   45.005   45.005 main.py:10(factorial)
    2    0.000    0.000    0.000    0.000 {built-in method builtins.print}
    1    0.000    0.000    0.000    0.000 {method 'disable' of '_lsprof.Profiler' objects}
    1    0.000    0.000    0.000    0.000 main.py:3(add)

Scalene

• ... is also cool
• scalene main.py

Visualization

• tuna
• snakeviz

Example

• python -m cProfile -o slow_random.prof slow_random.py
• snakeviz slow_random.prof
• go to web browser and see the results
• resulting plots will look like this:

... or this:

• each function is represented as a rectangle, with the size of the rectangle proportional to the time spent in that function. You can set given rectangle as root and see the time spent in it's sub-functions

References

• quick yt tutorial
• some tutorial
• daily dev list
• scalene
• snakeviz