[Python] Performance difference while calling similar, very simple functions in Python

I'm confused. I defined two silly functions:

def test_mul(r,f):
for i in range(1000):
x=r*8+f

def test_add(r,f):
for i in range(1000):
x=r+f


test_add is some 20% faster than test_mul, and that sounds reasonable (just under 17 secs for 1 million repetitions vs just under 22 secs on my notebook). Now let's use global variables instead of arguments

r=3
f=5

def test_mul2():
for i in range(1000):
x=r*8+f

def test_add2():
for i in range(1000):
x=r+f


These are incredibly faster, and moreover, test_mul2, which performs a multiplication and a sum, is consistently slightly faster than test_add2, which only performs a sum: around 11 millisecs vs around 12.

timeit('test_add(3,5)',globals=globals())
16.795306700048968
timeit('test_mul(3,5)',globals=globals())
21.756636400008574

timeit('test_add2',globals=globals())
0.012141100014559925
timeit('test_mul2',globals=globals())
0.010826400015503168


Can someone explain what's happening (Py 3.12 on Win 11)?

Continue reading...