Introduction Last year I wrote a series of posts (Part 1, Part 2, Part 3) on the use of the new Java fork/join framework for a Monte Carlo simulation. First, an update. Going back through the code I discovered a bug in the way the triangle distribution was implemented. Fortunately this is a toy example, as it made the results inaccurate. My fault for not unit testing. I would still not suggest using this code for anything other than learning about fork/join.

Introduction In a previous post I discussed the fork/join framework introduced with Java SE 7 and how it can be used to perform simple parallelism of certain types of tasks; that is, those that operate within a single JVM and involve a large piece of work that can be broken up into smaller pieces through a divide and conquer strategy. To illustrate this, I introduced an example of using fork/join to perform a Monte Carlo simulation of the Net Present Value of an investment with uncertain profits and discount rate.

This post continues a series discussing the new fork/join features added to Java SE 7. Part 1 and part 2 introduced an example application that performs a Monte Carlo simulation of the Net Present Value of an investment with uncertain profits and discount rate. The example application is available on GitHub. The previous post glossed over an important question: what is the “right” way to divide up the work? Of course, the answer is dependent on the type of work being performed and the machine on which it is run.

Introduction This is the first of a series of posts that will discuss an example application that uses the Java Fork/Join framework to do a Monte Carlo simulation of the Net Present Value of a prospective investment. This first post discusses the purpose of the example application as well as the purpose of the fork/join framework and the kinds of problems it can best solve. Fork/Join and Divide and Conquer The fork/join framework introduced with Java SE 7 is based on a specific model of parallel programming, similar to what is implemented in languages such as Intel’s Cilk Plus.