Unit 38: Threads

Learning Objectives

Students should

• understand the behaviour of thread.
• be able to create a simple thread.

Synchronous Programming

So far, when we invoke a method in Java, we expect the method to return us a value when it is done. If the method is not done, the execution of our program stalls, waiting for the method to complete its execution. Only after the method returns can the execution of our program continue.

Or, if we use the lazy monad, the execution is only done when we really need the value. But this will also stalls the execution of our program until the value is produced.

We say that the method blocks until it returns. Such a programming model is known as synchronous programming.

Synchronous programming is not very efficient, especially when there are frequent method calls that block for a long period (such as methods that involve expensive computations or reading from a remote server over the Internet).

What if we want our program to do something while we wait for the method to return? For instance, refreshing the UI, or do other computations?

Threads

One way to achieve this is to use threads. A thread is a single flow of execution in a program. Since the beginning of this module, we have been writing single-thread programs, except for parallel streams in Unit 37.

Java provides a class called java.lang.Thread that we can use to encapsulate a function to run in a separate thread. The following example how we can create and run two threads:

new Thread(() -> {
  for (int i = 1; i < 100; i += 1) {
    System.out.print("_");
  }
}).start();

new Thread(() -> {
  for (int i = 2; i < 100; i += 1) {
    System.out.print("*");
  }
}).start();

The new Thread(..) is our usual constructor to create a Thread instance. The constructor takes a Runnable instance as an argument. A Runnable is a functional interface with a method run() that takes in no parameter and returns void.

All Kinds of Functional Abstraction

At this point, we have all possible functional abstraction if we only consider functions with at most one parameter and at most one return value.

Functional Abstraction	Has Parameter?	Has Return Value?
Function<T, R>
Producer<T>
Consumer<T>
Runnable

If we consider multiple parameters, then there are infinitely many of these. But as we know from Currying, these are equivalent to a method with a single parameter.

With each Thread instance, we run start(), which causes the given lambda expression to run. Note that start() returns immediately. It does not return only after the given lambda expression completes its execution. This property differs from what we are used to, where a method blocks until the task given completes. This is known as asynchronous execution.

The two threads above now run in two separate sequences of execution. The operating system has a scheduler that decides which threads to run when, and on which core (or which processor). You might see different interleaving of executions every time you run the same program.

Java provides more than one way to create a thread. The Thread class also contains methods that we can use to query and control, in a finer grain manner, how the thread could be executed.

Names

Every thread in Java has a name, printing out its name is useful for peeking under the hood to see what is happening. We can use the instance method getName() to find out the name of a thread, and the class method Thread.currentThread() to get the reference of the current running thread.

System.out.println(Thread.currentThread().getName()); // main

new Thread(() -> {
  System.out.print(Thread.currentThread().getName());
  for (int i = 1; i < 100; i += 1) {
    System.out.print("_");
  }
}).start();

new Thread(() -> {
  System.out.print(Thread.currentThread().getName());
  for (int i = 2; i < 100; i += 1) {
    System.out.print("*");
  }
}).start();

Note that the above will also print the name of the thread called main, which is a thread created automatically for us every time our program runs and the class method main() is invoked.

With this method, you can now "visualize" how many parallel threads are created when you invoke a parallel stream.

Try

Stream.of(1, 2, 3, 4)
      .parallel()
      .reduce(0, (x, y) -> { 
        System.out.println(Thread.currentThread().getName()); 
        return x + y; 
      });

and you will see something like this (may not be exactly this as it depends on the worker thread available):

main
ForkJoinPool.commonPool-worker-5
ForkJoinPool.commonPool-worker-5
ForkJoinPool.commonPool-worker-9
ForkJoinPool.commonPool-worker-3
ForkJoinPool.commonPool-worker-3
ForkJoinPool.commonPool-worker-3

being printed. This shows four concurrent threads running to reduce the stream of 1, 2, 3, 4 (including main).

If you remove the parallel() call as shown below

Stream.of(1, 2, 3, 4)
      .reduce(0, (x, y) -> { 
        System.out.println(Thread.currentThread().getName()); 
        return x + y; 
      });

then only main is printed, showing the reduction being done sequentially in a single thread.

main
main
main
main

Sleep

Another useful method in the Thread class is Thread::sleep. You can cause the current execution thread to pause execution immediately for a given period (in milliseconds). After the sleep timer is over, the thread is ready to be chosen by the scheduler to run again.

The following code prints a "." on-screen every second while another expensive computation is running.

Thread findPrime = new Thread(() -> {
  System.out.println(
      Stream.iterate(2, i -> i + 1)
          .filter(i -> isPrime(i))
          .limit(1_000_000L)
          .reduce((x, y) -> y)
          .orElse(null));
});

findPrime.start();

while (findPrime.isAlive()) {
  try {
    Thread.sleep(1000);
    System.out.print(".");
  } catch (InterruptedException e) {
    System.out.print("interrupted");
  }
} 

In our examples, we often use Thread.sleep() in our methods to pretend that we are working hard on expensive computation to keep our examples simple.

Two more things to note:

• The example above shows how we use isAlive() to periodically check if another thread is still running.
• The program exits only after all the threads created run to their completion.