I tried it your way, and I got "No such file" and "Permission denied". pdftohtml generates its output in the current working directory. I believe that if you try any other directory, it'll ignore the command and generate it in the current directory anyway. I tried it this way:

#!/usr/bin/perl

use strict;
use warnings;

open(STDOUT, '>', 'htmllog');

my $cmd = 'pdftohtml -noframes -c /path/to/.pdf'; 
system($cmd);

exit;
[download]

Note that I wanted to keep a log to double check the number of pages.