Bignum operations slower than in Python

[Redysz]

Hi, I'm working with some big numbers operations. In other cases v is extremely faster than Python and close enough for me to C. However, with those big number operations Python is faster than v and it is crushing difference. First - powers:

import time
import math { powi }
import math.big

fn main() {
	start := time.now()
	a := big.integer_from_int(9)
	result := a.pow(33420489)
	stop := time.now()
	println("time elapsed: ${(stop-start).seconds()}")
}

On my computer it took 227.2s. Python's equivalent:

from time import perf_counter

def main():
    start = perf_counter()
    result = 9**33420489
    stop = perf_counter()
    print(f"time elapsed: {(stop-start):.2f}")

this took only 39.75s.
I've checked v's pow implementation for bigints and I couldn't find place for optimalization. It works as it should (using (x^2)^(n/2) "trick"). So... I checked multiplication big numbers:

import time
import math { powi }
import math.big

fn main() {
    a := big.integer_from_int(9)
	first := a.pow(99999)
	second := a.pow(88888)
	start := time.now()
	for _ in 0..1000 {
		result := first * second
	}
	stop := time.now()
	println("time elapsed: ${(stop-start).seconds()}")
}

it took 57.8s, and equivalent for Python:

from time import perf_counter

def main():
    first = 9**99999
    second = 9**88888
    start = perf_counter()
    for _ in range(1000):
        result = first * second
    stop = perf_counter()
    print(f"time elapsed: {(stop-start):.2f}")

took only 12s. Doing the same for + operation I can conclude that bigints in Python are just "somehow" implemented with better performance.
I can transform it to feature request or if that's where it should go. I would like to make some "big computation" in v.