How should i reduce the execution time using Threading in python?

Question

when i used the following code using single thread time is 49.7 secs but as i increase the number of threads as below

work1=TestThread(self,"worker1") 
work2=TestThread(self,"worker2")

the execution time is also increasing.

import time
import wx
import threading
from threading import Thread

# Define notification event for thread completion
EVT_RESULT_ID = wx.NewId()

def EVT_RESULT(win, func):
    """Define Result Event."""
    win.Connect(-1, -1, EVT_RESULT_ID, func)

class ResultEvent(wx.PyEvent):
    """Simple event to carry arbitrary result data."""
    def __init__(self, data):
        """Init Result Event."""
        wx.PyEvent.__init__(self)
        self.SetEventType(EVT_RESULT_ID)
        self.data = data

########################################################################
class TestThread(Thread):
    """Test Worker Thread Class."""

    #----------------------------------------------------------------------
    def __init__(self, wxObject,name):
        """Init Worker Thread Class."""
        Thread.__init__(self)
        self.name = name
        self.wxObject = wxObject
        self.start()    # start the thread

    #----------------------------------------------------------------------
    def run(self):
        """Run Worker Thread."""
        # This is the code executing in the new thread.
        start_time=time.time()
        print " Entering " + self.name + "\n"
        for i in range(1000):
            for a in range(1000):
                for b in range(1000):
                    pass
        print " execution time of "+ self.name +" is: " + str(time.time()-start_time) + "\n"
        aot=1
        wx.PostEvent(self.wxObject, ResultEvent(aot))

########################################################################
class MyForm(wx.Frame):

    #----------------------------------------------------------------------
    def __init__(self):
        wx.Frame.__init__(self, None, wx.ID_ANY, "Tutorial")

        # Add a panel so it looks the correct on all platforms
        panel = wx.Panel(self, wx.ID_ANY)
        self.displayLbl = wx.StaticText(panel, label="Amount of time since thread started goes here")
        self.btn = btn = wx.Button(panel, label="Start Thread")

        btn.Bind(wx.EVT_BUTTON, self.onButton)

        sizer = wx.BoxSizer(wx.VERTICAL)
        sizer.Add(self.displayLbl, 0, wx.ALL|wx.CENTER, 5)
        sizer.Add(btn, 0, wx.ALL|wx.CENTER, 5)
        panel.SetSizer(sizer)

        # Set up event handler for any worker thread results
        EVT_RESULT(self, self.updateDisplay)

    #----------------------------------------------------------------------
    def onButton(self, event):
        """
        Runs the thread
        """
        self.displayLbl.SetLabel("Thread started!")
        work1=TestThread(self,"worker1")
        print threading.enumerate()
        print threading.activeCount()
        btn = event.GetEventObject()
        btn.Disable()

    #----------------------------------------------------------------------
    def updateDisplay(self, msg):
        """
        Receives data from thread and updates the display
        """
        t = msg.data
        if (t==1):
            self.btn.Enable()

#----------------------------------------------------------------------
# Run the program
if __name__ == "__main__":
    app = wx.App()
    frame = MyForm().Show()
    app.MainLoop()

The output for single thread is:

Entering worker1
[<_MainThread(MainThread, started 9360)>, <Thread(SockThread, started daemon 4984)>, <TestThread(worker1, started 5776)>]

3
 execution time of worker1 is: 49.2760000229

for two thread's is:

Entering worker1
 Entering worker2
[<_MainThread(MainThread, started 8880)>, <Thread(SockThread, started daemon 8280)>, <TestThread(worker1, started 5788)>, <TestThread(worker2, started 12240)>]


4
 execution time of worker2 is: 113.527999878

 execution time of worker1 is: 114.292000055

Why is it happening like that. Can anyone explain? Also what is that SockThread? TIA

CPU bound Python programs will not see performance boost from multithreading due to the GIL. — thebjorn, Apr 30 '16 at 13:57
should i use both multiprocessing and multithreading together or only multiprocessing? — dinece, May 01 '16 at 14:12

score 1 · Answer 1 · answered Apr 30 '16 at 14:34

1

Your threads are fighting with eachother on one core due to the Global Interpreter Lock. The Global Interpreter Lock prevents more than one python thread from running at a time, even on a multi-core system. Thus, to execute multiple cpu-bound tasks in parallel, it is necessary to use processes (from the multiprocessing module) instead of threads.

Also, the SockThread appearing in your list is internal to IDLE.

answered Apr 30 '16 at 14:34

pppery

3,731
22
33
46

oh thank you. will try multiprocessing then and see. – dinece May 01 '16 at 11:18
should i use both multiprocessing and multithreading together or only multiprocessing alone ? – dinece May 02 '16 at 06:33
There is no reason to use the `threading` module for multiple cpu-bound tasks. It is just slower. Only use `multiprocessing` – pppery May 02 '16 at 12:59
Every python module is documented online. The documentation for `multiprocessing` is located at https://docs.python.org/2/library/multiprocessing.html – pppery May 11 '16 at 21:07
Thank You. I've seen this documentation but it wasn't clearly documented.Thats why i asked for a good documentation. :) – dinece May 12 '16 at 04:18
In the sense multiprocessing documentation related to wxpython – dinece May 12 '16 at 04:41

How should i reduce the execution time using Threading in python?

1 Answers1