It depends on the output encoding you intend to use for that file. If the encoding is utf-8, then the anonymous post above is correct, because length will return the number of bytes in UTF-8 when use bytes is in effect. But if your encoding is something else, the result may be different:

use strict;
use warnings;
my $x = "abcdefg";
print length $x, "\n";
{
  use bytes;
  print length $x, "\n";
}
open my $fh, ">:encoding(ucs-2)", "out.txt" or die $!;
print $fh $x;
close $fh;
print +(stat "out.txt")[7], "\n"; #print size of file in bytes
__END__
7
7
14
[download]

A more generic technique for determining the length of a string in a given output encoding:

use Encode qw/encode/;
my $x = "abcdefg";
print length(encode("ucs-2", $x)), "\n";
[download]