Java多種方法實現等待所有子執行緒完成再繼續執行

南瓜慢說發表於2023-03-08
Java執行緒

簡介

在現實世界中，我們常常需要等待其它任務完成，才能繼續執行下一步。Java實現等待子執行緒完成再繼續執行的方式很多。我們來一一檢視一下。

Thread的join方法

該方法是Thread提供的方法，呼叫join()時，會阻塞主執行緒，等該Thread完成才會繼續執行，程式碼如下：

private static void threadJoin() {
  List<Thread> threads = new ArrayList<>();

  for (int i = 0; i < NUM; i++) {
    Thread t = new Thread(new PkslowTask("Task " + i));
    t.start();
    threads.add(t);
  }
  threads.forEach(t -> {
    try {
      t.join();
    } catch (InterruptedException e) {
      throw new RuntimeException(e);
    }
  });

  System.out.println("threadJoin Finished All Tasks...");

}

結果：

Task 6 is running
Task 9 is running
Task 3 is running
Task 4 is running
Task 7 is running
Task 0 is running
Task 2 is running
Task 1 is running
Task 5 is running
Task 8 is running
Task 1 is completed
Task 8 is completed
Task 6 is completed
Task 4 is completed
Task 3 is completed
Task 0 is completed
Task 7 is completed
Task 9 is completed
Task 2 is completed
Task 5 is completed
threadJoin Finished All Tasks...

CountDownLatch

CountDownLatch是一個很好用的併發工具，初始化時要指定執行緒數，如10。在子執行緒呼叫countDown()時計數減1。直到為0時，await()方法才不會阻塞。程式碼如下：

private static void countDownLatch() {
  CountDownLatch latch = new CountDownLatch(NUM);
  for (int i = 0; i < NUM; i++) {
    Thread t = new Thread(() -> {
      System.out.println("countDownLatch running...");
      try {
        Thread.sleep(1000);
        System.out.println("countDownLatch Finished...");
        latch.countDown();
      } catch (InterruptedException e) {
        throw new RuntimeException(e);
      }
    });
    t.start();
  }

  try {
    latch.await();
  } catch (InterruptedException e) {
    throw new RuntimeException(e);
  }
  System.out.println("countDownLatch Finished All Tasks...");
}

結果：

countDownLatch running...
countDownLatch running...
countDownLatch running...
countDownLatch running...
countDownLatch running...
countDownLatch running...
countDownLatch running...
countDownLatch running...
countDownLatch running...
countDownLatch running...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished...
countDownLatch Finished All Tasks...

CyclicBarrier

CyclicBarrier與CountDownLatch類似，但CyclicBarrier可重置，可重用。程式碼如下：

private static void cyclicBarrier() {
  CyclicBarrier barrier = new CyclicBarrier(NUM + 1);

  for (int i = 0; i < NUM; i++) {
    Thread t = new Thread(() -> {
      System.out.println("cyclicBarrier running...");
      try {
        Thread.sleep(1000);
        System.out.println("cyclicBarrier Finished...");
        barrier.await();
      } catch (InterruptedException | BrokenBarrierException e) {
        throw new RuntimeException(e);
      }
    });
    t.start();
  }

  try {
    barrier.await();
  } catch (InterruptedException | BrokenBarrierException e) {
    throw new RuntimeException(e);
  }
  System.out.println("cyclicBarrier Finished All Tasks...");
}

結果：

cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier running...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished...
cyclicBarrier Finished All Tasks...

executorService.isTerminated()

ExecutorService呼叫shutdown()方法後，可以透過方法isTerminated()來判斷任務是否完成。程式碼如下：

private static void executeServiceIsTerminated() {
  ExecutorService executorService = Executors.newFixedThreadPool(THREADS);
  IntStream.range(0, NUM)
    .forEach(i -> executorService.execute(new PkslowTask("Task " + i)));
  executorService.shutdown();
  while (!executorService.isTerminated()) {
    //waiting...
  }
  System.out.println("executeServiceIsTerminated Finished All Tasks...");

}

結果：

Task 0 is running
Task 2 is running
Task 1 is running
Task 3 is running
Task 4 is running
Task 0 is completed
Task 2 is completed
Task 5 is running
Task 4 is completed
Task 7 is running
Task 3 is completed
Task 1 is completed
Task 8 is running
Task 6 is running
Task 9 is running
Task 5 is completed
Task 9 is completed
Task 7 is completed
Task 6 is completed
Task 8 is completed
executeServiceIsTerminated Finished All Tasks...

executorService.awaitTermination

executorService.awaitTermination方法會等待任務完成，並給一個超時時間，程式碼如下：

private static void executeServiceAwaitTermination() {
  ExecutorService executorService = Executors.newFixedThreadPool(THREADS);
  IntStream.range(0, NUM)
    .forEach(i -> executorService.execute(new PkslowTask("Task " + i)));
  executorService.shutdown();

  try {
    if (!executorService.awaitTermination(1, TimeUnit.MINUTES)) {
      executorService.shutdownNow();
    }
  } catch (InterruptedException e) {
    throw new RuntimeException(e);
  }
  System.out.println("executeServiceAwaitTermination Finished All Tasks...");
}

結果：

Task 0 is running
Task 1 is running
Task 2 is running
Task 3 is running
Task 4 is running
Task 0 is completed
Task 5 is running
Task 1 is completed
Task 4 is completed
Task 7 is running
Task 3 is completed
Task 8 is running
Task 2 is completed
Task 9 is running
Task 6 is running
Task 5 is completed
Task 7 is completed
Task 9 is completed
Task 8 is completed
Task 6 is completed
executeServiceAwaitTermination Finished All Tasks...

executorService.invokeAll

使用invokeAll提交所有任務，程式碼如下：

private static void executeServiceInvokeAll() {
  ExecutorService executorService = Executors.newFixedThreadPool(THREADS);
  List<Callable<Void>> tasks = new ArrayList<>();

  IntStream.range(0, NUM)
    .forEach(i -> tasks.add(new PkslowTask("Task " + i)));

  try {
    executorService.invokeAll(tasks);
  } catch (InterruptedException e) {
    throw new RuntimeException(e);
  }

  executorService.shutdown();
  System.out.println("executeServiceInvokeAll Finished All Tasks...");
}

結果：

Task 1 is running
Task 2 is running
Task 0 is running
Task 3 is running
Task 4 is running
Task 1 is completed
Task 3 is completed
Task 0 is completed
Task 2 is completed
Task 4 is completed
Task 8 is running
Task 5 is running
Task 6 is running
Task 9 is running
Task 7 is running
Task 8 is completed
Task 5 is completed
Task 6 is completed
Task 9 is completed
Task 7 is completed
executeServiceInvokeAll Finished All Tasks...

ExecutorCompletionService

ExecutorCompletionService透過take()方法，會返回最早完成的任務，程式碼如下：

private static void executorCompletionService() {
  ExecutorService executorService = Executors.newFixedThreadPool(10);
  CompletionService<String> service = new ExecutorCompletionService<>(executorService);

  List<Callable<String>> callables = new ArrayList<>();
  callables.add(new DelayedCallable(2000, "2000ms"));
  callables.add(new DelayedCallable(1500, "1500ms"));
  callables.add(new DelayedCallable(6000, "6000ms"));
  callables.add(new DelayedCallable(2500, "2500ms"));
  callables.add(new DelayedCallable(300, "300ms"));
  callables.add(new DelayedCallable(3000, "3000ms"));
  callables.add(new DelayedCallable(1100, "1100ms"));
  callables.add(new DelayedCallable(100, "100ms"));
  callables.add(new DelayedCallable(100, "100ms"));
  callables.add(new DelayedCallable(100, "100ms"));

  callables.forEach(service::submit);

  for (int i = 0; i < NUM; i++) {
    try {
      Future<String> future = service.take();
      System.out.println(future.get() + " task is completed");
    } catch (InterruptedException | ExecutionException e) {
      throw new RuntimeException(e);
    }
  }

  System.out.println("executorCompletionService Finished All Tasks...");

  executorService.shutdown();
  awaitTerminationAfterShutdown(executorService);
}

這裡不同任務的時長是不一樣的，但會先返回最早完成的任務：

2000ms is running
2500ms is running
300ms is running
1500ms is running
6000ms is running
3000ms is running
1100ms is running
100ms is running
100ms is running
100ms is running
100ms is completed
100ms is completed
100ms task is completed
100ms task is completed
100ms is completed
100ms task is completed
300ms is completed
300ms task is completed
1100ms is completed
1100ms task is completed
1500ms is completed
1500ms task is completed
2000ms is completed
2000ms task is completed
2500ms is completed
2500ms task is completed
3000ms is completed
3000ms task is completed
6000ms is completed
6000ms task is completed
executorCompletionService Finished All Tasks...

程式碼

程式碼請看GitHub: https://github.com/LarryDpk/pkslow-samples

相關文章