Executor is a simple standardized interface for defining custom thread-like subsystems, including thread pools, asynchronous I/O, and lightweight task frameworks. Depending on which concrete Executor class is being used, tasks may execute in a newly created thread, an existing task-execution thread, or the thread calling
execute, and may execute sequentially or concurrently.
ExecutorService provides a more complete asynchronous task execution framework. An ExecutorService manages queuing and scheduling of tasks, and allows controlled shutdown. The
ScheduledExecutorService subinterface and associated interfaces add support for delayed and periodic task execution. ExecutorServices provide methods arranging asynchronous execution of any function expressed as
Callable, the result-bearing analog of
Future returns the results of a function, allows determination of whether execution has completed, and provides a means to cancel execution. A
RunnableFuture is a
Future that possesses a
run method that upon execution, sets its results.
ScheduledThreadPoolExecutor provide tunable, flexible thread pools. The
Executors class provides factory methods for the most common kinds and configurations of Executors, as well as a few utility methods for using them. Other utilities based on
Executors include the concrete class
FutureTask providing a common extensible implementation of Futures, and
ExecutorCompletionService, that assists in coordinating the processing of groups of asynchronous tasks.
ForkJoinPool provides an Executor primarily designed for processing instances of
ForkJoinTask and its subclasses. These classes employ a work-stealing scheduler that attains high throughput for tasks conforming to restrictions that often hold in computation-intensive parallel processing.
Five implementations in
java.util.concurrent support the extended
BlockingQueue interface, that defines blocking versions of put and take:
DelayQueue. The different classes cover the most common usage contexts for producer-consumer, messaging, parallel tasking, and related concurrent designs.
TransferQueue, and implementation
LinkedTransferQueue introduce a synchronous
transfer method (along with related features) in which a producer may optionally block awaiting its consumer.
Interfaces and classes providing a framework for locking and waiting for conditions that is distinct from built-in synchronization and monitors.
Condition factors out the
Object monitor methods (
notifyAll) into distinct objects to give the effect of having multiple wait-sets per object, by combining them with the use of arbitrary
Lock implementations provide more extensive locking operations than can be obtained using
synchronized methods and statements.
TimeUnit class provides multiple granularities (including nanoseconds) for specifying and controlling time-out based operations. Most classes in the package contain operations based on time-outs in addition to indefinite waits. In all cases that time-outs are used, the time-out specifies the minimum time that the method should wait before indicating that it timed-out. Implementations make a "best effort" to detect time-outs as soon as possible after they occur. However, an indefinite amount of time may elapse between a time-out being detected and a thread actually executing again after that time-out. All methods that accept timeout parameters treat values less than or equal to zero to mean not to wait at all. To wait "forever", you can use a value of
Five classes aid common special-purpose synchronization idioms.
Semaphoreis a classic concurrency tool.
CountDownLatchis a very simple yet very common utility for blocking until a given number of signals, events, or conditions hold.
CyclicBarrieris a resettable multiway synchronization point useful in some styles of parallel programming.
Phaserprovides a more flexible form of barrier that may be used to control phased computation among multiple threads.
Exchangerallows two threads to exchange objects at a rendezvous point, and is useful in several pipeline designs.
Besides Queues, this package supplies Collection implementations designed for use in multithreaded contexts:
CopyOnWriteArraySet. When many threads are expected to access a given collection, a
ConcurrentHashMap is normally preferable to a synchronized
HashMap, and a
ConcurrentSkipListMap is normally preferable to a synchronized
CopyOnWriteArrayList is preferable to a synchronized
ArrayList when the expected number of reads and traversals greatly outnumber the number of updates to a list.
The "Concurrent" prefix used with some classes in this package is a shorthand indicating several differences from similar "synchronized" classes. For example
Collections.synchronizedMap(new HashMap()) are synchronized. But
ConcurrentHashMap is "concurrent". A concurrent collection is thread-safe, but not governed by a single exclusion lock. In the particular case of ConcurrentHashMap, it safely permits any number of concurrent reads as well as a tunable number of concurrent writes. "Synchronized" classes can be useful when you need to prevent all access to a collection via a single lock, at the expense of poorer scalability. In other cases in which multiple threads are expected to access a common collection, "concurrent" versions are normally preferable. And unsynchronized collections are preferable when either collections are unshared, or are accessible only when holding other locks.
Most concurrent Collection implementations (including most Queues) also differ from the usual
java.util conventions in that their Iterators and Spliterators provide weakly consistent rather than fast-fail traversal:
9.Memory Consistency Properties
Chapter 17 of the Java Language Specification defines the happens-before relation on memory operations such as reads and writes of shared variables. The results of a write by one thread are guaranteed to be visible to a read by another thread only if the write operation happens-before the read operation. The
volatile constructs, as well as the
Thread.join() methods, can form happens-before relationships. In particular:
synchronizedblock or method exit) of a monitor happens-before every subsequent lock (
synchronizedblock or method entry) of that same monitor. And because the happens-before relation is transitive, all actions of a thread prior to unlocking happen-before all actions subsequent to any thread locking that monitor.
volatilefield happens-before every subsequent read of that same field. Writes and reads of
volatilefields have similar memory consistency effects as entering and exiting monitors, but do not entail mutual exclusion locking.
starton a thread happens-before any action in the started thread.
joinon that thread.
The methods of all classes in
java.util.concurrent and its subpackages extend these guarantees to higher-level synchronization. In particular:
Executorhappen-before its execution begins. Similarly for
Callablessubmitted to an
Futurehappen-before actions subsequent to the retrieval of the result via
Future.get()in another thread.
CountDownLatch.countDownhappen-before actions subsequent to a successful "acquiring" method such as
CountDownLatch.awaiton the same synchronizer object in another thread.
Exchanger, actions prior to the
exchange()in each thread happen-before those subsequent to the corresponding
exchange()in another thread.
Phaser.awaitAdvance(as well as its variants) happen-before actions performed by the barrier action, and actions performed by the barrier action happen-before actions subsequent to a successful return from the corresponding
awaitin other threads.