import random
import sys

def estimate_pi(n: int) -> float:
    num_points_in_circle = 0 
    num_points_total = n
    for i in range(n):
        x = random.uniform(0,1)
        y = random.uniform(0,1)
        if x**2 + y**2 <= 1:
            num_points_in_circle += 1
    return 4 * num_points_in_circle / num_points_total

def main():
    input = int(sys.argv[1])
    print(estimate_pi(input))

main()

will you get the same time usage using above script compared with your python code? I got 5s on my compute.

And you know, Python have a new compiler called codon, it is a high-performance Python compiler that compiles Python code to native machine code without any runtime overhead.

you can try to compile the above script into native code by running:

codon build --release -o calpi calpi.codon

after that, run time ./calpi 10000000 to have a try . I got a result of 0.18s on my compute.