Python multiprocessing

@lawrenceyy wrote:

I just started exploring multiprocessing in python outside of ghpythonlib.parallel.run(). I've run into some unexpected results.

My test script:

import multiprocessing as mp
import time
from PIL import Image as image

class myclass:
    def __init__(self,integer):
        self.integer=integer
        self.result=None
        self.image=None

def func(_index):
    #this function uses PIL to generate a solid color image
    #the purpose is to cause significant CPU load
    myclasslist[_index].image=image.new("L",(2000,2000),0)
    temp=[myclasslist[_index].integer*20 for k in range(2000*2000)]
    myclasslist[_index].image.putdata(temp)
    #myclasslist[_index].result is what I am testing
    myclasslist[_index].result=myclasslist[_index].integer+100
    print (myclasslist[_index].result)
    return myclasslist[_index].result

time0=time.time()

myclasslist=[myclass(i) for i in range(10)]
pool=mp.Pool(processes=4)
mpresult=pool.map(func,range(10))

time1=time.time()

print(time1-time0)
print (mpresults)
print([myclasslist[i].result for i in range(10)])

In the end mpresults show the correct results, but myclasslist[i].result shows None. I read online that each process gets its own instance of my object, therefore when I do:
myclasslist[_index].result=myclasslist[_index].integer+100
it does not stick. Is there a way to get it to stick so when I print in the end it shows the correct values?

And does anyone know how to get pool.map() to work with class methods()?

Posts: 1

Participants: 1

Read full topic