Tasks in Parallel

Resources:

• Tatiana Fesenko (O’Reilly): Video

ForkJoin Framework

Problem: It is hard to execute tasks in parallel because we don’t know how many threads to create, and coordinating their communication is error prone.
Solution: Using ForkJoin can simplify by managing a pool of threads and by synchronizing the results. A common pool allows the framework to optimize based on available resources.
How: Create a collection of tasks (or Callables). Use a ForkJoinPool to optimally invoke the collection on separate threads and to join the results.

Callable

Callable<V>: A Generic Functional Interface. The method has no arguments and returns something of type V.

Callable<Integer> c1 = () -> { logger.info(10); return 10; };

The ForkJoin Framework can execute a list of Callables on separate threads and return a list of Futures.

List<Future<Integer>> futures = pool.invokeAll(myCallables);

Future