A function for the min definition necessary for parallel operations with python's multiprocessing

parallel_multiprocess.py

import multiprocessing

# Parallel
# So, it goes like this:
# Given the number of 'to process' items,
# Check how the distribution for cores will be (eg: 100 items on 2 cores = 50)
# Then, split items in *number of cores* collections and spread them aroung a
# subprocess pool
# To split the set, we create the interval [0, number_of_items] and make it
# go up the original items collection *number of cores* times
# [Return] tuple with : 'map appliable' pool of subprocesses
#                       'splitted' dataset
def split_dataset(data, set_size, cores):
  splitted_dataset = []
  for i in range(0, cores):
    splitted_dataset.append(data[i*set_size:(i+1)*set_size])
  return splitted_dataset

def parallelize(data):
  number_of_cores = multiprocessing.cpu_count()
  set_size = len(data) / number_of_cores
  splitted = split_dataset(data, set_size, number_of_cores)
  pool = multiprocessing.Pool(number_of_cores)
  return (pool, splitted)

# For testes
import time
import random

def func(numb):
  print "processing %s" % numb
  val = random.randint(1, numb[0] + 2)
  time.sleep(int(val))
  print "ended processing %s" % numb

data = list(range(0,100))

pool, dataset = parallelize(data)
pool.map(func, dataset)