Parallelizing loop in python

I have an agents population of 10,000 and each agent has its location [x,y], where x and y are random numbers between 0 and 100. I want to calculate distance between each agent and decide if they are neighbours and then store these information in 10000x10000 array. So far I've come up with the following:

for j in range(len(agents_population[1])):
    B[j,j:10000] = [
        True if distance.euclidean(agents_population[1][j],
            agents_population[1][i]) < r 
        else False for i in range(j,len(agents_population[1]))
      ]

Where agents_population[1] is a list of agents' coordinates (so in this case 10000-items list of 2-items lists: [[x1,y1], [x2,y2]...), r is radius and B is that array. In each iteration I calculate distance between agent j and agents j+1 to 9999. However, each iteration of the loop is independent from each other, so I was thinking I could use all four cores of my processor instead of just one. I tried using multiprocessing module and Pool().map() function, but I wasn't able to make it work. I would be really grateful for any advice.

EDIT:

I tried using multiprocessin as follows:

def worker(j):
    B[j,j:10000] = [
        True if distance.euclidean(agents_population[1][j],
            agents_population[1][i]) < r 
        else False for i in range(j,len(agents_population[1]))
      ]

def mp_handler():
    p = Pool(4)
    p.map(worker, range(0))

if __name__ == '__main__':
    mp_handler() 

B[0,].tofile('foo.csv',sep=',')

I run it via cmd.exe, but when I open file foo.csv the results are different compared to when I run function worker with j=0. Am I doing something wrong?