gistfile1.txt

#!/usr/bin/env perl

use strict;
use warnings;

use Benchmark qw(cmpthese);

cmpthese shift || -3, {
    "shift"     => sub {
        my $arg = shift;
    },
    "array"     => sub {
        my($arg) = @_;
    },
    "1 element scalar" => sub {
        my $arg = $_[0];
    },
    "1 element list" => sub {
        my($arg) = $_[0];
    },
};


__DATA__
NOTE: This is benchmarking a subroutine being passed NO arguments.
It could all change if it were passed arguments.
Also these numbers are fairly volatile, changing as much as 20% between runs.
Not surprising as we're benchmarking such a fast operation. 

5.10.1 no threads
                       Rate      array 1 element list     shift 1 element scalar
array            11904762/s         --           -15%      -35%             -48%
1 element list   13986014/s        17%             --      -23%             -39%
shift            18181818/s        53%            30%        --             -21%
1 element scalar 22988506/s        93%            64%       26%               --

5.12.4 with threads
                       Rate      array 1 element list     shift 1 element scalar
array            11764706/s         --            -5%      -28%             -45%
1 element list   12422360/s         6%             --      -24%             -42%
shift            16260163/s        38%            31%        --             -24%
1 element scalar 21276596/s        81%            71%       31%               --

5.14.1 with threads
                       Rate      array 1 element list     shift 1 element scalar
array             9433962/s         --           -17%      -36%             -49%
1 element list   11363636/s        20%             --      -23%             -38%
shift            14705882/s        56%            29%        --             -20%
1 element scalar 18348624/s        94%            61%       25%               --

5.16.1 no threads
                       Rate      array 1 element list 1 element scalar     shift
array            10638298/s         --           -13%             -56%      -61%
1 element list   12195122/s        15%             --             -49%      -55%
1 element scalar 24096386/s       127%            98%               --      -11%
shift            27027027/s       154%           122%              12%        --