I`m working on a project where I want to do analysis,
but the analysis is quite intense and would like to exploit all my CPU`s.

I looked into multi threading, first option was (IMHO) the map function of python.
But that function seems not to exploit all my CPUs..

So I found the following post: http://python.rhino3d.com/entries/60...hreaded-Python
Multi threading works great with that script UNTIL I use an numpy object.

Example:
Code:
##nessecary lines for Rhino&Scipy/Numpy
import clr
clr.AddReference("mtrand")
#Import and set Rhino
import numpy as np

#Get the bridge to Rhino
import rhinoscriptsyntax as rs
import Rhino
import System.Drawing
import scriptcontext
import System.Threading.Tasks as tasks
import time


big = 1000
A = np.array( [[1,1,1],[1,1,1],[1,1,1]] )
print "Test map function"

def function(value):
    
    temp = value
    for i in range(big):
        temp = temp + value
        B = np.dot(A,A)
    


start = time.time() 
for i in range(big):
     function(i) 
print time.time() - start

start = time.time()
map(function, range(big))
print time.time() - start

start = time.time() 
tasks.Parallel.ForEach(range(big), function)
print time.time() - start
Could someone tell me how I could multithread the numpy module?

Kind regards!