Let's say I need to spawn multiple threads to do the work, and continue to the next step only after all of them complete. I will need to tell the main thread to wait. The key point is to use Thread.join() method. For example,

package foo;
import java.util.Vector;
public class ThreadTest {
private Vector<String> threadNames = new Vector<String>();
public static void main(String[] args) {
ThreadTest test = new ThreadTest();
test.threadTest(Integer.parseInt(args[0]));
System.out.println(test.threadNames);
}
private void threadTest(int numOfThreads) {
Thread[] threads = new Thread[numOfThreads];
for (int i = 0; i < threads.length; i++) {
    threads[i] = new foo.ThreadTest.MyThread();
    threads[i].start();
}
for (int i = 0; i < threads.length; i++) {
    try {
       threads[i].join();
    } catch (InterruptedException ignore) {}
}
}
class MyThread extends Thread {
 public void run() {
     for (int i = 0; i < 1000000; i++) {
         i = i + 0;
     }
     threadNames.add(getName());
 }
}
}

The output when running this program with 10 threads:

[Thread-1, Thread-3, Thread-0, Thread-5, Thread-7, Thread-6, Thread-9, Thread-2, Thread-8, Thread-4]

The order in which the threads are executed is random, which is expected.

Also note that we use two for-loops, the first to create and start each thread, and the second loop to join each thread. If each thread is joined right after start, the effect is these threads are executed sequentially, without the desired concurrency. For example, the following code snippet results in serial execution:

// The following is NOT concurrent:
//
private void threadTest(int numOfThreads) {
 Thread[] threads = new Thread[numOfThreads];
 for (int i = 0; i < threads.length; i++) {
     threads[i] = new foo.ThreadTest.MyThread();
     threads[i].start();
     try {
         threads[i].join();
     } catch (InterruptedException ignore) {
     }
 }
}

Output:

[Thread-0, Thread-1, Thread-2, Thread-3, Thread-4, Thread-5, Thread-6, Thread-7, Thread-8, Thread-9]

If we don't use any join at all, threadNames, when printed, may be empty, or partially filled, since the main thread will just move on when it gets the chance. The main thread will still wait, at the very last step, for all threads to complete, before exiting the JVM. The output for running 10 threads may be:

[Thread-0, Thread-1, Thread-2, Thread-3]

The following is a general-purpose code skeleton for running concurrency test using anonymous Runnable inner-class:

        Thread[] threads = new Thread[numOfThreads];
        for (int i = 0; i < threads.length; i++) {
            threads[i] = new Thread(new Runnable() {
                public void run() {
                    System.out.println("xxx");
                }
            });
            threads[i].start();
        }
        for (int i = 0; i < threads.length; i++) {
            try {
                threads[i].join();
            } catch (InterruptedException e) {
            }
        }

Another version using java.util.concurrent.CountDownLatch: When to join threads with CountDownLatch