Here's a quick comparison of PDF-related modules

Wait, I don't see any beyond (incomplete) enumeration, looks like you forgot to append results. Oh, then, here they are (some of them for a start), and using simple test file generated with code you kindly provided:

use strict;
use warnings;
use PDF::API2;
use CAM::PDF;
use Benchmark 'cmpthese';

my $fn = 'test.pdf';
my $str = do { local ( @ARGV, $/ ) = $fn; <> };

cmpthese -1, {
  'PDF::API2' => sub { PDF::API2-> from_string( $str )-> page_count },
  'CAM::PDF'  => sub { CAM::PDF-> new( $str )-> numPages },
  'CAM::PDF+' => sub { my $d = CAM::PDF-> new( $str ); 
                       $d-> cacheObjects; $d-> numPages },
};

__END__

            Rate PDF::API2 CAM::PDF+  CAM::PDF
PDF::API2  163/s        --      -74%      -95%
CAM::PDF+  614/s      277%        --      -83%
CAM::PDF  3586/s     2106%      484%        --
[download]

The 'plussed' entry (kind of "parse everything") is for those who may have (reasonable) doubts if perhaps one module (guess which) makes harder effort to extract a lot more info initially, to provide a user with richer environment to inspect things more cosily (or something like that); but in fact, they (nonplussed) seem both to extract approximately same amount of info on open. It's just one parser (guess which) is very poor indeed.