Benchmark - Slots ou pas slots ?

Benchmark pour évaluer l’intérêt de définir un slot dans une classe Python:

import timeit
from sys import getsizeof


class SlottedPoint:
    __slots__ = ("x", "y")

    def __init__(self, x, y):
        self.x = x
        self.y = y



class NonSlottedPoint:

    def __init__(self, x, y):
        self.x = x
        self.y = y


def dictionary():
    return {
        "x": 1,
        "y": 2
    }

def check_func_performance(func: str, call_count: int = 1000000, repeat_count: int = 10):
    func_time = timeit.repeat(func, number=call_count, globals=globals(), repeat=repeat_count)
    func_time_value = sum(func_time) / repeat_count
    size = getsizeof(eval(func))

    return {
        "title": func,
        "duration": func_time_value,
        "object_size": f"{size} Byte(s)"
    }



results = check_func_performance('SlottedPoint(1, 2)')
print(results)

results = check_func_performance('NonSlottedPoint(1, 2)')
print(results)

results = check_func_performance('dictionary()')
print(results)

Résultats - Python 3.9:

func	duration	size
SlottedPoint(1, 2)	0.3355669799999987	48 Byte(s)
NonSlottedPoint(1, 2)	0.42160326000000625	48 Byte(s)
dictionary()	0.19480267000001278	232 Byte(s)

Résultats - Python 3.10:

func	duration	size
SlottedPoint(1, 2)	0.5916475000325591	48 Byte(s)
NonSlottedPoint(1, 2)	0.5538422700483352	48 Byte(s)
dictionary()	0.26400925999041647	232 Byte(s)