Python Tutorial - 26. Multithreading - Introduction

Hello friends, in today's code basics coding tutorial, we are going to talk about multi-threading in Python. Before we begin with Python multi-threading, let's first understand in general what multi-threading means when it comes to computer science. Multi-threading I often explain uh by giving the example of a busy mom. Uh as you see in this picture, busy mom usually juggles between multiple tasks. So she might be cooking food and at the same time she has to take care of baby and she often gets a phone call and she needs to handle that as well in parallel. So as you see on the right hand side, mom starts by cooking food and at some point you see in this in the green line she has to switch to another task of taking care of uh a baby and then at some point here uh she gets a phone call so she needs to handle that as well. So in this time period, so this vertical slice indicates the time progression. In this time period, she's cooking food, taking care of baby and uh handling phone call. So each of these lines are individual tasks that mom is handling. And when it comes to programming, they are called threads. So mom is handling three threads. Cooking food, taking care of baby and handling phone call. Hence mom is doing multi-threading. Similarly, Python program can handle multiple tasks at the same time and that program uh is called to be doing multi-threading. Okay. So we going to work on a problem where I have an array a list of numbers and I want to calculate the square of each of these numbers and I want to print those numbers at the same time. I want to calculate cube of each of these numbers and print them. So I have to handle two task basically calculating squares and calculating numbers calculating cubes of these numbers in the array. uh so it's simple to write two functions so what I have written here is I have written the first function which is calculating square which takes numbers array as an input I go through the array and I print the square here now I'm introducing a time sleep so this time module what it does is it waits for 2 seconds before printing this number and I will tell you in a bit why I'm doing this But for now just assume that this goes through an input array and it just prints square of each of these numbers. This function is exactly similar to this function except that instead of square it is calculating the cube of those numbers. Okay. And I'm calculating I'm calling calculating square and calculate cube one by one. And then I'm printing done when I'm done. And I'm also doing something else which is uh calculating how much time it took to execute these two functions. Okay. So it's fairly simple program. Uh so let let me just run it. Okay. So as you can see here I have this array 2 389. It first printed 4 9 64 81 uh the square of these numbers and the cube of these numbers. So cube of 2 is 2 into 2 into 2 is 8. 3 cube is 27 because 3 into 3 9 into 3 uh is equal to 27 and so on. Overall this program took 1.6 second. Let me execute one more one more time. Again it took 1.6 seconds. So note down this number. this program is taking 1.6 second to execute. Okay. Now the reason I introduced this time uh delay here is that just to demonstrate in which scenario multi-threading could be useful. Here you are first calculating square and then calculating cube. Now when you're waiting here what's happening is your CPU is idle. See your CPU is doing nothing. Okay. Now you might have this scenario when let's say in real life you might be calling a web service or you might be waiting for a packet on your network socket. So at that time you are waiting and your CPU in your PC is not doing anything. Multi-threading tries to utilize this idle time and during that idle time you want to give CPU some work. Okay. So let's use multi-threading and improve this number. So it's taking right now 1.6 second. Using multi-threading we will improve the time will execute this program much faster. Okay. So first thing you need to do in order to use multi-threading is import threading module. Okay. This is a standard Python module used for multi-threading. And now instead of calling these fun functions directly, so I'm going to remove these. And what I'm going to do is I'm going to create a thread. So I'll say threading uh threading dot thread. Okay. and target equal to. So your target is your worker function the the task simple task that you want to execute. This is similar to mom taking a phone call for example or or handling baby. So in our case that task is calculating square. Okay. So you mention your function name after target and in your args you will mention your arguments. So here your arguments is your array. So this is a tpple that you're passing as an input. So if you have multiple arguments to this function, you will pass multiple arguments here. Okay. So this way you created the first thread and now I'm going to create the second thread and here my function will be calculating cube. Okay. So by doing this I'm creating these two threads T1 and T2. Now I just created a threads two threads and I need to uh start them now. So I will say T1 dot start then T2 dot start. So this will execute these two programs in parallel. Okay. And then you need to say T1.join join T2 dot join what join will do is it will wait until these this thread is done basically T1 dot join means now wait here until t1 is done t1 meaning until this uh this calculate square function is finished and then you you want to print done okay so let's execute this program Now nice. So as you see here now what it is doing is it is not executing these two functions in a sequence. So first is it calculated a square. Okay. So it calculated a square here. Then it went into this loop one more time and it and then it started waiting. So while it was waiting here the other function was also executing in parallel. So meanwhile it printed cube. So you see cube here. So that's why you see square and cube because there are two tasks being performed simultaneously in parallel. Okay. So whichever task gets done you get the output of that here. And you see that now our program is taking half the time compared to what it used to take before. So here is8. If you recall previously it used to take 1.6 second. Now it is taking only.8 second. Okay. Now let's compare this with our busy mom example. So here the main program started. So this red line is called the main thread. Then at some point we started calculating square of numbers. So which is this green line. So we just spawned a new thread which is this green color line. So that we did by doing this T1 start. So when you say T1 start what you're doing is you are just basically start starting execution of this new thread. Then at some point you started calculating Q of numbers which is this D2 dot start and then when the they are done they come back to this main program. So these points here are basically T1.join and T2.join. Join means you are done with your usual thread. Now the execution joins back into your main program your main thread. Okay. So again comparing it with bimom is kind of very similar thing that your program doing. Okay. So multi-threading is typically used when as explained previously when your CPU is idle it is waiting for something. I didn't have a sophisticated example that's why I introduce artificial delay. But in in your real life you'll be calling let's say web service. web service might take few seconds to give you back the response during that time. If you want to optimize the performance of your program and you make your CPU do some work then you can definitely use uh multi-threading. Now multi-threading in Python is a little special because there is something called a glo global interpreter lock that uh prevents you from using the true benefits of multi-threading. But in any case whenever you are waiting and when you are doing IO bound operation you can still use multi-threading. So go ahead and use it. And if you want to do real CPU in intensive work and you're not waiting for an IIO type operation then you need to use multiprocessing. So all those are little advanced concepts which we will cover in future videos. Okay. So that was a sweet little introduction on multi-threading. Thank you very much for watching.