I am using a computer with many cores and for performance benefits I should really use more than one. However, I'm confused why these bits of code don't do what I expect:
from multiprocessing import Process
var = range(5)
def test_func(i):
global var
var[i] += 1
if __name__ == '__main__':
jobs = []
for i in xrange(5):
p = Process(target=test_func,args=(i,))
jobs.append(p)
p.start()
print var
As well as
from multiprocessing import Pool
var = range(5)
def test_func(i):
global var
var[i] += 1
if __name__ == '__main__':
p = Pool()
for i in xrange(5):
p.apply_async(test_func,[i])
print var
I expect the result to be [1, 2, 3, 4, 5] but the result is [0, 1, 2, 3, 4].
There must be some subtlety I'm missing in using global variables with processes. Is this even the way to go or should I avoid trying to change a variable in this manner?
解决方案
If you are running two separate processes, then they won't be sharing the same globals. If you want to pass the data between the processes, look at using send and recv. Take a look at http://docs.python.org/library/multiprocessing.html#sharing-state-between-processes for an example similar to what you're doing.