Skip to content

Instantly share code, notes, and snippets.

@csivanich
Last active August 29, 2015 14:01
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save csivanich/1b6210920d37255a7c43 to your computer and use it in GitHub Desktop.
Save csivanich/1b6210920d37255a7c43 to your computer and use it in GitHub Desktop.
Java Concurrency

Java Concurrency

docs.oracle.com


Process

  • Self contained execution environment
  • Own memory space
  • Interprocess execution environment
  • Process have at least one

Threading Strategies

  • Instantiate own Threads on demand
  • Pass tasks to an Executor (As described later)

Defining and starting a thread

  • Two methods:
    • Best practice: define Runnable class object
    • Other option: Extend Thread and execute Thread.run()

Pausing Threads

Thread.sleep()

  • Suspends thread for defined time
  • Not guaranteed to be precise. It's as precise as the OS may be
  • Cannot assume it will not be interrupted

Interrupts

  • Usually given with InterruptedException
  • Thread should stop and do something else
  • Very common for stopping threads

Join

  • Waits for a Thread to complete
  • Can have timeout like Thread.sleep() throwing InterruptedException when timed out
public class SimpleThreads {

    // Display a message, preceded by
    // the name of the current thread
    static void threadMessage(String message) {
        String threadName =
            Thread.currentThread().getName();
        System.out.format("%s: %s%n",
                          threadName,
                          message);
    }

    private static class MessageLoop
        implements Runnable {
        public void run() {
            String importantInfo[] = {
                "Mares eat oats",
                "Does eat oats",
                "Little lambs eat ivy",
                "A kid will eat ivy too"
            };
            try {
                for (int i = 0;
                     i < importantInfo.length;
                     i++) {
                    // Pause for 4 seconds
                    Thread.sleep(4000);
                    // Print a message
                    threadMessage(importantInfo[i]);
                }
            } catch (InterruptedException e) {
                threadMessage("I wasn't done!");
            }
        }
    }

    public static void main(String args[])
        throws InterruptedException {

        // Delay, in milliseconds before
        // we interrupt MessageLoop
        // thread (default one hour).
        long patience = 1000 * 60 * 60;

        // If command line argument
        // present, gives patience
        // in seconds.
        if (args.length > 0) {
            try {
                patience = Long.parseLong(args[0]) * 1000;
            } catch (NumberFormatException e) {
                System.err.println("Argument must be an integer.");
                System.exit(1);
            }
        }

        threadMessage("Starting MessageLoop thread");
        long startTime = System.currentTimeMillis();
        Thread t = new Thread(new MessageLoop());
        t.start();

        threadMessage("Waiting for MessageLoop thread to finish");
        // loop until MessageLoop
        // thread exits
        while (t.isAlive()) {
            threadMessage("Still waiting...");
            // Wait maximum of 1 second
            // for MessageLoop thread
            // to finish.
            t.join(1000);
            if (((System.currentTimeMillis() - startTime) > patience)
                  && t.isAlive()) {
                threadMessage("Tired of waiting!");
                t.interrupt();
                // Shouldn't be long now
                // -- wait indefinitely
                t.join();
            }
        }
        threadMessage("Finally!");
    }
}
  • Locks done through internal Java API known as monitors
  • Allows only one thread to hold the lock
  • 2 threads accessing at once causes one to block until other gives up the monitor
  • Automatic acquisition and release when using synchronize'd methods
  • Static methods provide class level locks

Synchronized Statements

  • Allows Object to be specified as the intrinsic lock Example using lock1 and lock2 as locks:
public class MsLunch {
    private long c1 = 0;
    private long c2 = 0;
    private Object lock1 = new Object();
    private Object lock2 = new Object();

    public void inc1() {
        synchronized(lock1) {
            c1++;
        }
    }

    public void inc2() {
        synchronized(lock2) {
            c2++;
        }
    }
}
  • Must be careful with using this method
  • Make sure interleaving access to the fields is alright

Atomic Access

  • References and primitives are atomic writes
    • Except long and double
  • Reads are atomic for all volatile variables
  • Atomic access is more efficient than synchronize'd code
  • It requires more care, however

Deadlocks

  • Two Threads waiting for each other to do something
  • Threads will never end and must be interrupted

Starvation

  • synchronize'd method which is slow
  • One Thread calls method many times
  • Another Thread must now wait long periods before using method

Livelock

  • Two Threads constantly responding to eachother
  • Not technically blocked, simply too busy to respond

Guarded Blocks

  • Coordinated actions between Threads
  • Works by polling a condition that must be true to continue
  • Uses Object.wait() until interrupted and condition check can try again

Example:

public synchronized void guardedJoy(){
    while ( ! joy ){
        try {
            wait();
        } catch ( InterruptedException e ) {}
    }
    System.out.println("We have joy!");
}
  • Note the InterruptedException is not the end of the loop, it just works to start the Thread back up and check the conditional once more
  • Invoking wait() inside synchronize'd methods is an easy way to acquire the intrinsic lock
  • Object.notifyAll() is usually better than Object.notify()
  • Object.notify() does not specify which Thread is woken up
  • Java Collections Framework has guarded block objects

High Level Concurrency Objects

Lock Objects - Locking idioms to simplify concurrent applications

Executors - API for thread management

Concurrent Collections - Collections with reduced need of synchronize

Atomic Variables - Minimize syncing + memory consistency issues

ThreadLocalRandom - Psudorandom numbers from multiple Threads

Lock Objects and the Lock Interface

  • Very similar to implicit locks
  • Support wait/notify through condition objects
  • Advantages:
  • Can back out of lock acquisition attempt
  • Lock.tryLock() backs out if not available or times out
  • Lock.lockInterruptibly() backs out if another Thread interrupts prior to acquisition

Executors - Interfaces, Pools and Fork/Join

  • 3 Interfaces:
  • Executor - Launcher for new tasks
  • ExecutorService - Subinterface of Executor with lifecycle management features
  • ScheduledExecutorService - Scheduled subinterface of ExecutorService

Executor Interface

  • Executor.execute(Runnable)
  • The java.util.concurrent implementation uses worker threads and more advanced functionality such as thread pools

ExecutorService Interface

  • ExecutorService.submit(Runnable || Callable)
  • Returns Future object
  • Retrieves Callable return value
  • Manages status of Callable and Runnable tasks
  • Methods for submitting Collections of Callable tasks
  • Manages shutdown of the Executor

ScheduledExecutorService Interface

  • Executes Runnable or Callable tasks after a delay
  • ScheduledExecutorService.scheduleAtFixedRate()
  • ScheduledExecutorService.scheduleWithFixedDelay()

Thread Pools - java.util.concurrent.Executors

  • Minimizes Thread objects to instantiate
  • Alloc/Dealloc of Thread memory space is expensive
  • Fixed Thread Pool:
  • Specified number of Threads
  • Terminated Threads are replaced with new ones
  • Submitted via internal queue for when there are more tasks than available worker Threads
  • Graceful degredation
  • Can't spin up more Threads
  • Keeps execution within system limits, even when many tasks

Fork & Join

  • Implementation of ExecutorService
  • Built for multicore/many processors
  • Designed for small units of work built recursively
  • Goal is to use all available hardware power
  • Uses worker threads in a thread pool
  • Idle worker threads steal tasks from other busy threads
Basic Use
  • ForkJoinTask
  • Usually RecursiveTask or RecursiveAction Impl
if ( work is small enough )
    do it all
else
    split the work
    invoke this on each piece and wait for results

Written with StackEdit.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment