Java SE 8 parallelSort in Arrays class
Arrays class now has several new “parallel” methods to take advantage of multi-core hardware. Arrays method parallelSort can sort large arrays more efficiently on multicore systems. Lets create a very large array and use Java SE 8 Date/Time API to compare how long it takes to sort the array with methods sort and parallelSort.
SecureRandom random = new SecureRandom();
// generate array of random ints, then copy it intArray1
int[] intArray1 = random.ints(18_000_000).toArray();
// Create another arrary
int[] intArray2 = new int[intArray1.length];
System.arraycopy(intArray1, 0, intArray2, 0, intArray1.length);
System.out.println("Perform Arrays.sort");
Instant sortStart = Instant.now();
Arrays.sort(intArray1);
Instant sortEnd = Instant.now();
long sortTime = Duration.between(sortStart, sortEnd).toMillis();
System.out.printf("Total time in milliseconds: %d%n%n", sortTime);
// time the sorting of intArray2 with Arrays method parallelSort
System.out.println("Perform Arrays.parallelSort");
Instant parallelSortStart = Instant.now();
Arrays.parallelSort(intArray2);
Instant parallelSortEnd = Instant.now();
long parallelSortTime =Duration.between(parallelSortStart, parallelSortEnd).toMillis();
System.out.printf("Total time in milliseconds: %d%n%n",parallelSortTime);
String percentage = NumberFormat.getPercentInstance().format((double) sortTime / parallelSortTime);
System.out.printf("%nArrays.sort took %s more time than Arrays.parallelSort%n", percentage);
Output
Perform Arrays.sort Total time in milliseconds: 1738 Perform Arrays.parallelSort Total time in milliseconds: 435 Arrays.sort took 400% more time than Arrays.parallelSort