How to benefit from Java 8's parallelStream?

Question

I asked a question about dynamic filtering using Java 8 at [ How to dynamically do filtering in Java 8? ], Stuart Marks answered it wonderfully and gave some sample code to test with at : https://gist.github.com/stuart-marks/10076102.

Since I'm new to Java 8 and so I wonder if the following lines would make the program run faster on multi-core workstations with lots of data to process, at least theoretically :

//  widgetList.stream().filter(compositePredicate).forEach(System.out::println);
    widgetList.parallelStream().filter(compositePredicate).forEach(System.out::println);

//  compositeCriteria.apply(widgetList.stream()).forEach(System.out::println);
    compositeCriteria.apply(widgetList.parallelStream()).forEach(System.out::println);

Would relacing "widgetList.stream()" with "widgetList.parallelStream()" speedup the process ?

Answer 1

The best way would be to test it. However note that:

• System.our.println is a synchronized method, so the terminal operation is essentially sequential
• only the filtering can therefore benefit from the parallelisation
• a parallel stream will probably be a gain for fairly large lists or for computationally intensive filtering operations