1
00:00:05,800 --> 00:00:08,560
I hope you enjoyed this section on concurrency.

2
00:00:08,830 --> 00:00:13,330
We will now be going over the solution to the assignment for this section.

3
00:00:13,420 --> 00:00:17,740
The question was using Rayon for this solution research.

4
00:00:17,770 --> 00:00:18,850
Rayon join.

5
00:00:19,180 --> 00:00:24,790
Once you have read into rayon, join in the Fibonacci join function of the provided source code.

6
00:00:24,790 --> 00:00:31,450
Create a solution that utilizes rayon join so that we can see a performance enhancement in our Fibonacci

7
00:00:31,450 --> 00:00:32,230
solution.

8
00:00:33,400 --> 00:00:42,340
So once you research rayon join, you took away some information where rayon join takes two closures

9
00:00:42,340 --> 00:00:45,940
and then possibly runs them in parallel.

10
00:00:45,940 --> 00:00:56,950
In this case, they will be ran in parallel and then join is going to return a pair of the results from

11
00:00:56,980 --> 00:00:58,390
those closures.

12
00:00:59,290 --> 00:01:05,860
So we know that we're going to get some values back from the two closures.

13
00:01:05,860 --> 00:01:14,260
So we'll create a tuple that is going to store values A and B and we're going to call Rayon join.

14
00:01:15,640 --> 00:01:19,030
And we know that Rayon join once two closures.

15
00:01:19,030 --> 00:01:25,540
So we'll give it this closure where we're going to call the Fibonacci recursive function.

16
00:01:25,540 --> 00:01:30,940
And in here we're going to pass in N minus one.

17
00:01:31,240 --> 00:01:40,030
And now we know that we need to pass in another closure and this time we will pass in N minus two.

18
00:01:41,620 --> 00:01:49,780
And then simply all we have to do is add A plus B, so now we can run this and we see that in our first

19
00:01:49,780 --> 00:02:00,190
iteration of fib recursive, which is the non parallel ized function, we see that Fibonacci recursive

20
00:02:00,190 --> 00:02:07,780
actually takes a decent bit of time to actually execute because what it's doing is it's creating our

21
00:02:07,780 --> 00:02:14,770
series of numbers for us and it happens to be pretty intensive computation.

22
00:02:14,770 --> 00:02:22,630
So we see we get a really large number here and it took us 18 and one half seconds, but when we ran

23
00:02:22,630 --> 00:02:30,970
it with Range join therefore introducing some parallelism, we were able to knock off a little over

24
00:02:30,970 --> 00:02:35,650
6 seconds and still generate the correct solution.

25
00:02:36,010 --> 00:02:44,230
So hopefully this is an eye opening assignment on how much time can be saved when you introduce some

26
00:02:44,230 --> 00:02:46,810
parallelism into your solutions.

27
00:02:46,810 --> 00:02:52,720
So I hope this assignment was fun for you and I will see you in the next section.