2
\$\begingroup\$

I have a synchronized list that I use as a taskqueue. In this queue I have some long values that need to be proceed and create heavy I/O and delay. I thought of using threads to get things done faster.

Is there a better way to implement all 50 threads to wait than using t.join();? Is there anything wrong with the way I access the queue in my run() implementation?

Implementation:

public class Start {

    private static List<Long> queue = Collections.synchronizedList(new ArrayList<Long>());
    private static HashSet<MyThread> threads = new HashSet<MyThread>();


    public static void main(String[] args) throws InterruptedException {
        // add elements to workqueue
        for(int i = 0; i < 100000; i++) {
            queue.add(new Date().getTime());
        }

        // create 50 new threads
        for (int i = 0; i < 50; i++) {
             MyThread t = new MyThread(queue);
             t.start();
             threads.add(t);
        }

        // wait for all 50 threads to finish
        for (MyThread t : threads) {
            t.join();
        }

        System.out.println("Done");
    }
}

Thread:

public class MyThread extends Thread {
    private List<Long> queue = null;

    public MyThread(List<Long> queue) {
        this.queue = queue;
    }

    @Override
    public void run() {
        boolean _break = false;
        int size = 0;
        long task = -1;
        while(true) {
            synchronized (queue) {
                if(!queue.isEmpty()) {
                    size = queue.size();
                    // get task for worker
                    task = queue.remove(0);
                } else {
                    _break = true;
                }
            }
            if(_break == true) {
                System.out.println("Thread-" + this.getId() + " Size: 0");
                break;
            } else {
                // do some stuff with the task
                // ...
                // ...
                System.out.println("Thread-" + this.getId() + " Size: " + size);
            }
        }
    }
}
\$\endgroup\$

1 Answer 1

1
\$\begingroup\$

Use an Executor

Maintaining a work queue and a thread pool are problems that an executor tackles:

This interface provides a way of decoupling task submission from the mechanics of how each task will be run, including details of thread use, scheduling, etc. An Executor is normally used instead of explicitly creating threads.

The use of Futures also helps you impose time constraints and handle exceptions thrown during execution.

(Note that you'll typically want to use ExecutorService or CompletionService as Executor is a minimalistic interface.)

To paraphrase your example:

class IncrediblyLongRunningTask implements Callable<Long> {
  @Override
  public Long call() throws Exception {
    // ...
  }
}

public void compute() throws InterruptedException {  
  final ExecutorService service = Executors.newFixedThreadPool(50);
  final List<IncrediblyLongRunningTask> tasks = new ArrayList<>();
  for ( int i = 0; i < /*...*/; i++ ) {
    task.add(new IncrediblyLongRunningTask());
  }
  final List<Future<Long>> futures = service.invokeAll(tasks); // returns when all done
  service.shutdown();
  for ( final Future<Long> future : futures ) {
    // ...
  }
}
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.