The reality of today's chip architectures s that rather than seeing ever increasing clock speeds we now see ever increasing core counts. This has a significant impact on how we write efficient code since we now need to design for parallel operations to get the best from our hardware. Concurrent programming is hard to get right, especially when faced with early Java support in the form of the sleep, interrupt, wait and notify methods along with synchronised blocks and methods with a single monitor per object. Java SE 5 started the process of enhancing Java's support for parallel programming with the concurrency APIs. Java SE 7 includes the latest revisions to this in the form of the fork-join framework (also known as JSR-166y).

This session will look at how the fork-join framework can greatly simplify the code required to process a set of tasks that can be distributed among a number of parallel processing units. We'll look at some good examples for effective use of the framework as well as one that is not suited and provide some hints on what to do in this case.