perf_plot

A Python performance measuring and visualization tool using gnuplot.

Example Usage:

from perf_plot import timeit, perf_test
import sys

sys.dont_write_bytecode = 1  # Include this to prevent pycache

d = dict()


@timeit
def key_test(num: int):
    vals = d.values()


def make_dict(num: int):
    d = {key: key for key in range(num)}


@timeit
def unpack_test(num: int):
    vals = [*d]


perf_test(
    [
        unpack_test,
        key_test,
    ],
    begin=10**3,
    bound=10**5,
    step=10**3,
    callbacks=[make_dict],
)

Prerequisites

You must have gnuplot installed to use this tool.

API

perf_test function

def perf_test(funcs: list, begin=1, bound=10000, step=100, callbacks=[]):

Run a list of functions with a range of inputs, time them, and plot their performance.

list funcs

A list of tester functions that:

• Each take an int, and
• Do some comparable operations that scale accordingly

e.g [labmda x: [i for i in range(x)], lamba y: (i for i in range(y))]

int begin

The lower bound of the test, defaults to 1

int bound

The upper bound of the test, defaults to 10000

int step

How much to increment for each iteration, defaults to 100

list callbacks

A list of callbacks to run before testing each function, defaults to []

@timeit decorator

@timeit
def example_func(n):  # Replace with a function you want to time
    return n * 2

@timeit decorator: Time a function