Doug Lea: A Java Fork/Join Framework, Proceedings of the ACM 2000 conference on Java Grande.
This paper describes the design, implementation, and performance of a Java framework for supporting a style of parallel programming in which problems are solved by (recursively) splitting them into subtasks that are solved in parallel, waiting for them to complete, and then composing results. The general design is a variant of the workâˆ’stealing framework devised for Cilk.
This work is about to be incorporated into Java 7 as jsr166y:
Parallel*Array (often referred to as PA) and its planned follow-ons for sets and maps, provide an easier/better way of routinely programming to take advantage of dozens to hundreds of processors/cores: If you can think about a programming problem in terms of aggregate operations on collections of elements, then we can automate parallel execution. This generally pays off if either you have lots of elements, (in which case, it works well even if the operations are small/cheap), or if each of the operations are time consuming (in which case it works well even if there are not a lot of elements). To take advantage of this though, the aggregate processing must have a regular structure, which means that you must be able to express things in terms of apply, reduce, filter, map, cumulate, sort, uniquify, paired mappings, and so on.