Your point is taken, so I reran the tests and got this,

              Rate grep_loop  for_loop
grep_loop 150375/s        --      -22%
for_loop  193618/s       29%        --
[download]

So now they are both faster, but grep is still slower. Than the for loop

And as usual grep EXPR, LIST that everyone forgets about wins the race.

#!/usr/bin/perl
use strict;
use warnings;

my @values = qw(test test1 test2 test3);

use Benchmark qw( cmpthese );

cmpthese -2, {
    for_loop  => sub { my @matches; for (@values) { push @matches, $_ 
+if $_ =~ /test/ }; },
    grep_loop => sub { my @matches = grep { $_ =~ /test/ } @values;   
+                   },
    grep_expr => sub { my @matches = grep $_ =~ /test/, @values       
+                   },
};
__END__
              Rate grep_loop  for_loop grep_expr
grep_loop 248507/s        --      -27%      -46%
for_loop  338839/s       36%        --      -27%
grep_expr 461521/s       86%       36%        --
[download]

You owe the oracle a Battle Royale With Cheese and a tasty drink to wash it down with.

The cake is a lie.
The cake is a lie.
The cake is a lie.

I had always gotten into the habit of using {} with grep, that I did not think about dropping them.

This was one of the reasons I started this thread, was to gain some insight and learn a new approaches to make improvements to my code.

That was very useful bit of information, thanks!