List::Compare を使ったリストの差分抽出……なんでこんなに速いんだろう？

test.t

use strict;
use warnings;
use t::Utils;
use Test::More;

use List::MoreUtils qw/natatime/;
use List::Compare;
use Math::Random::MT;
use Time::Seconds;
use Time::HiRes qw/gettimeofday tv_interval/;

my $random = Math::Random::MT->new;

my $list1 = [];
my $list2 = [];

for (my $i = 1; $i <= 1000000; $i++) {
    push @$list1, sprintf("%7d", $i);
    push @$list2, sprintf("%7d", $i);
}

my @expect;
my $n = 1000000;
for (0..99999) {
    my $i = int $random->rand($n--);
    push @expect, splice @$list2, $i, 1;
}
diag "list1 num: " . scalar(@$list1);
diag "list2 num: " . scalar(@$list2);
diag "expect: ";
# diag explain @expect;

my $list3 = [];
for (@$list2) {
    my $x = int ( $_ / 5000);
    $x-- if $x == ($_ / 5000);
    $list3->[$x] = [] unless $list3->[$x];
    push @{$list3->[$x]}, $_;
}

subtest 'list matching' => sub {
    my $start = [gettimeofday];

    my @diff;
    # my @diff = grep { !{ map { $_, 1 } @$list2 }->{$_} } @$list1;

    # my $count = 0;
    # my $it = natatime(5000, @$list1);
    # while (my @parts = $it->()) {
    #     my $start_sub = [gettimeofday];
    #     my @parts_diff = grep { !{ map {$_, 1 } @{$list3->[$count]} }->{$_} } @parts;
    #     @diff = (@diff, @parts_diff);
    #     my $end_sub = [gettimeofday];
    #     diag "$count > ";
    #     diag explain @parts_diff;
    #     diag "sub time: " . tv_interval($start_sub, $end_sub);
    #     $count++;
    # }

    my $count = 0;
    my $it = natatime(5000, @$list1);
    while (my @parts = $it->()) {
        my $start_sub = [gettimeofday];
        my $lc = List::Compare->new(\@parts, $list3->[$count]);
        my @parts_diff = $lc->get_Lonly;
        @diff = (@diff, @parts_diff);
        my $end_sub = [gettimeofday];
        diag "$count";
        # diag explain \@parts_diff;
        diag "sub time: " . tv_interval($start_sub, $end_sub);
        $count++;
    }

    my $end = [gettimeofday];
    diag explain \@diff;
    diag "time: " . tv_interval($start, $end);
    is scalar(@diff), 100000;
    is_deeply [sort(@expect)], [sort(@diff)];
};

done_testing;