Data Parallelism - Java 8 Lambdas - page 99

If you can get away with using stateless operations, then you will get better parallel perform-
ance. Examples of stateless operations include map , filter , and flatMap ; sorted , dis-
tinct , and limit are stateful.

NOTE

Performance-test your own code. The advice in this section offers rules of thumb about
what performance characteristics should be investigated, but nothing beats measuring and
profiling.

Parallel Array Operations
Java 8 includes a couple of other parallel array operations that utilize lambda expressions
outside of the streams framework. Like the operations on the streams framework, these are
data parallel operations. Let's look at how we can use these operations to solve problems that
are hard to do in the streams framework.

These operations are all located on the utility class Arrays , which also contains a bunch of
other useful array-related functionality from previous Java versions. There is a summary in
Table 6-1 .

Table 6-1. Parallel operations on arrays

Name Operation

parallelPrefix Calculates running totals of the values of an array given an arbitrary function

parallelSetAll Updates the values in an array using a lambda expression

Sorts elements in parallel

parallelSort

You may have written code similar to Example 6-7 before, where you initialize an array us-
ing a for loop. In this case, we initialize every element to its index in the array.

Example 6-7. Initializing an array using a for loop
public public static static double double [] imperativeInitilize ( int int size ) {

double double [] values = new new double double [ size ];