A first benchmark

Let’s run and plot a first simple benchmark in Python.

Suppose I want to compare the efficiency of a few different methods to pre-allocate memory for a list of strings in Python.

Let’s define a separate python function for three different list pre-allocation strategies. These functions all take an integer called n as an input, which stands for the length of the list to be pre-allocated. The argument name n is the pyquickbench.default_ax_name and can be changed in as described in Preparing inputs.

def comprehension(n):
    return ['' for _ in range(n)]

def star_operator(n):
    return ['']*n

def for_loop_append(n):
    l = []
    for _ in range(n):
        l.append('')

all_funs = [
    comprehension   ,
    star_operator   ,
    for_loop_append ,
]

Let’s define relevant sizes of lists to be timed.

n_bench = 12
all_sizes = [2**n for n in range(n_bench)]

Now, let’s import pyquickbench, run and plot the benchmark

import pyquickbench

pyquickbench.run_benchmark(
    all_sizes   ,
    all_funs    ,
    show = True ,
)

From this benchmark, it is easy to see that initializing a list to a given length filled with a single value is quickest using the star operator.