[Python] Why Numpy fabs is much slower than abs?

This Python 3.12.7 script with Numpy 2.2.4 on AMD Ryzen 7 3800X:

import numpy as np, timeit as ti

a = np.random.rand(1000).astype(np.float32)
print(f'Minimum, median and maximum execution time in us:')

for fun in ('np.fabs(a)', 'np.abs(a)'):
t = 10**6 * np.array(ti.repeat(stmt=fun, setup=fun, globals=globals(), number=1, repeat=999))
print(f'{fun:20} {np.amin(t):8,.3f} {np.median(t):8,.3f} {np.amax(t):8,.3f}')


produces these results:

Minimum, median and maximum execution time in us:
np.fabs(a) 1.813 1.843 4.929
np.abs(a) 0.781 0.811 1.463


indicating that np.fabs() is more than 2x slower than np.abs(), despite the latter having more functionality. What is the reason?

Continue reading...