If you want to have a go at doing a highly parallel version of a simple task, or see real solutions, you could do worse than look at the wide finder project. Basically it's about how to do parallel regex matching of log files efficiently, but trying to add as little code as possible.
Participants have submitted solutions in many different languages and the performance results are posted. The original project has now finished and there is now wide finder 2 taking the work on.
CodingHorror has a good introduction to wide finder.
