How is faster the Lexical Subroutine

lexical-subs.pl

#!/usr/bin/env perl
# my $p = sub {} vs. my sub p {}
use strict;
use warnings;
use Benchmark ':all';
use Test::More 'no_plan';

sub sr1 {
    my $p = sub { return 2 * shift };
    return $p->(shift);
}

sub sr2 {
    my sub p { return 2 * shift };
    return p(shift);
}

is sr1(2), 4;
is sr2(2), 4;

printf("Running with Perl %s on %s\n%s\n", $^V, $^O, '-' x 80);
cmpthese(6e6, {
        'my $p = sub {}' => sub { sr1(2) },
        'my sub p {...}' => sub { sr2(2) },
    }
);

__END__
Running with Perl v5.30.0 on darwin
--------------------------------------------------------------------------------
                    Rate my sub p {...} my $p = sub {}
my sub p {...} 1197605/s             --           -76%
my $p = sub {} 4958678/s           314%             --

Running with Perl v5.32.0 on darwin
--------------------------------------------------------------------------------
                    Rate my sub p {...} my $p = sub {}
my sub p {...} 1315789/s             --           -73%
my $p = sub {} 4878049/s           271%             --