Sort_Test.pl - Sorting Performance Test

Sorting Algorithm Tutorials - Herong's Tutorial Examples

∟Sorting Algorithms Implementations in Perl

∟Sort_Test.pl - Sorting Performance Test

This section describes a sample test program, Sort_Test.pl, which can be used to test any sorting algorithm for data elements stored an array in Perl language.

If you like to program in Perl language, you may want to try to implement sorting algorithms in Perl. Here is my example of the sorting algorithm test program in Perl:

#- Sort_Test.pl
#- Copyright (c) 2015 HerongYang.com. All Rights Reserved.
#-
  use Time::HiRes qw(time);
  use Data::Dumper;
  do 'Sort_Function.pl';

  my ($arraySize, $numberOfTests) = @ARGV;
  $arraySize = 10 unless $arraySize;
  $numberOfTests = 1000 unless $numberOfTests;

  my $sortTime = 0;
  for (my $j=0; $j<$numberOfTests; $j++) {
    @dataArray = ();
    for (my $i=0; $i<$arraySize; $i++) {
      $dataArray[$i] = rand();
    }

    my $startTime = 1000*time();
    @dataArray = sort(@dataArray);
#    insertionSort(\@dataArray, 0, $arraySize);
#    selectionSort(\@dataArray, 0, $arraySize);
#    bubbleSort(\@dataArray, 0, $arraySize);
#    quickSort(\@dataArray, 0, $arraySize);
#    mergeSort(\@dataArray, 0, $arraySize);
#    heapSort(\@dataArray, 0, $arraySize);
#    shellSort(\@dataArray, 0, $arraySize);

#    print(Dumper(@dataArray));
    $sortTime += 1000*time() - $startTime;
  }
  $sortTime = $sortTime/$numberOfTests;

  print("Array size: $arraySize\n");
  print("Average sorting time: $sortTime milliseconds\n");
  print("Number of tests: $numberOfTests\n");
  printTimeInfo($sortTime, $arraySize);

sub printTimeInfo {
  my ($t, $n) = @_;
  my $nFactor = $t/($n/1000);
  my $nnFactor = $t/(($n/1000)*$n);
  my $log2n = log($n)/log(2.0);
  my $ngFactor = $t/(($n/1000)*$log2n);
  print("Performance: $nFactor O(N) microseconds\n");
  print("Performance: $ngFactor O(N*Log2(N)) microseconds\n");
  print("Performance: $nnFactor O(N*N) microseconds\n");
}

Note that:

I am repeating the sorting call 1000 times with randomly generated integers each time to compute the average sorting time.
The commented out function call statements represent implementations of different sorting algorithms provided in the included Sort_Functions.pl script file.
sort() is a Perl built-in implementation of the Quicksort algorithm. It is used to provide a performance baseline.
The Dumper() statement is provided just in case you want to verify if the array elements are really sorted.
The printTimeInfo() method shows how the implementation performs against 3 measuring scales: O(N), O(N*log2(N)), and O(N*N).
All variables are declared with "my" scope to avoid collision with variables in the sorting functions.

Before implementing my own versions of different sorting algorithms, I used the above test program to get some performance results with the built-in sort() function provided by Perl 5.18. Here are the results:

herong> perl -version
This is perl 5, version 18, subversion 2 (v5.18.2) built
for darwin-thread-multi-2level
Copyright 1987-2013, Larry Wall

herong> perl Sort_Test.pl 10000
  Array size: 10000
  Average sorting time: 10.6218154296875 milliseconds
  Number of tests: 1000
  Performance: 1.06218154296875 O(N) microseconds
  Performance: 0.0799371263185609 O(N*Log2(N)) microseconds
  Performance: 0.000106218154296875 O(N*N) microseconds

herong> perl Sort_Test.pl 20000
  Array size: 20000
  Average sorting time: 21.9370893554687 milliseconds
  Number of tests: 1000
  Performance: 1.09685446777344 O(N) microseconds
  Performance: 0.0767690753170121 O(N*Log2(N)) microseconds
  Performance: 5.48427233886719e-05 O(N*N) microseconds

herong> perl Sort_Test.pl 30000
  Array size: 30000
  Average sorting time: 36.1204841308594 milliseconds
  Number of tests: 1000
  Performance: 1.20401613769531 O(N) microseconds
  Performance: 0.0809549154666525 O(N*Log2(N)) microseconds
  Performance: 4.01338712565104e-05 O(N*N) microseconds

herong> perl Sort_Test.pl 100000
  Array size: 100000
  Average sorting time: 170.799684570313 milliseconds
  Number of tests: 1000
  Performance: 1.70799684570313 O(N) microseconds
  Performance: 0.102831656611221 O(N*Log2(N)) microseconds
  Performance: 1.70799684570313e-05 O(N*N) microseconds

The result was not very impressive comparing the Java and PHP languages. For example, the Java version took 25 milliseconds for array size of 100000, but the Perl version took 171 milliseconds for the same array size. So I lower the array size to 10000, 20000 and 30000 to build the performance baseline.

Array Size        10000   20000   30000   100000   200000   300000
----------        -----   -----   -----   ------   ------   ------
JDK Arrays.sort                               25       66      112
PHP sort()            3       7      13       75
Perl sort()          11      22      36      171