perl substr, verse substition, versus lvalue substring speed comparison using Benchmark. regular substr won by a larger margin than I thought, which makes me generally happy

output

# running 5.10.0, but according to http://perldoc.perl.org/perl5160delta.html#Performance-Enhancements, these benchmarks might be out of sync:
# Assignment to substr in void context is now more than twice its previous speed. Instead of creating and returning a special lvalue scalar that is then assigned to, substr modifies the original string itself.
# substr no longer calculates a value to return when called in void context.

Benchmark: running l_substr, sed, substring for at least 10 CPU seconds...
  l_substr: 11 wallclock secs (10.30 usr +  0.01 sys = 10.31 CPU) @ 1455470.71/s (n=15005903)
       sed: 10 wallclock secs (10.35 usr +  0.04 sys = 10.39 CPU) @ 433279.21/s (n=4501771)
 substring: 11 wallclock secs (10.44 usr +  0.01 sys = 10.45 CPU) @ 2926646.89/s (n=30583460)
               Rate       sed  l_substr substring
sed        433279/s        --      -70%      -85%
l_substr  1455471/s      236%        --      -50%
substring 2926647/s      575%      101%        --

substition-vs-substr-vs-lsubstr-perl-5.10.0.pl

#!/usr/bin/env perl
use strict;
use warnings;
use Benchmark qw( timethese cmpthese ) ;

my $r = timethese( -10, {
        sed => sub {
		my $sed_str = "This is a sentence";
		$sed_str =~ s/ is/ isnt/;
	},
	substring => sub {
		my $substring_str = "This is a sentence";
		substr($substring_str, 5, 2, "isnt");
	},
	l_substr => sub {
		my $l_str = "This is a sentence";
		substr($l_str, 5, 2) = "isnt";
	},
} );
cmpthese $r;