Dirty parallelisation of Python program

Question

I've got a program that loads 2600 images in a loop, does some processing and returns a value. Pseudo code:

for file in files:
  codes[file] = my_function(file)
return codes

The problem is - this process takes around 20-30 minutes, and seems to only use one CPU core. I'm looking for a quick and dirty way to run this across more cores, perhaps by splitting the list in two. I've read a bunch about Python and it's problems/solutions, but I can't figure what to do next. How do I do this?

FYI Using Python 2.7 on 2009 Macbook Pro (Core 2 Duo) w/ Numpy Scipy Scikit-image & OpenCV.

Answer 1

The easiest way is with multiprocessing:

from multiprocessing import Pool

with Pool() as p:
    all_codes = p.map(my_function, files)
return {f:code for code, f in zip(all_codes, files)}