Numba with right way

numba_vectorize.py

import timeit

statements = \
"""
import numpy as np
from numba import vectorize


def python_f(x,y):
  if x > y:
     m = x
  else:
     m = y
  return m

@vectorize
def numba_f(x,y):
  if x > y:
     m = x
  else:
     m = y
  return m

# your data, 10 million items
size = int(1e5)
x = np.random.randint(0, 20, size=size)
y = np.random.randint(0, 20, size=size)

numba_f(x,y)
"""


# pythonic way of element-wise operations
print(timeit.timeit('[python_f(x[i], y[i]) for i in range(len(x))]', setup=statements, number=500))

# numpy has its own vectorize
print(timeit.timeit('numpy_f = np.vectorize(python_f)', setup=statements, number=500))

# but numba's vectorize is what we're after
print(timeit.timeit('numba_f(x, y)', setup=statements, number=500))

inspired by https://roman-kh.github.io/numba-1/

results

• 22.113725809380412
• 0.0012139994651079178
• 0.05286153592169285