Assuming the *.md5files contain just the value (single line ?) then consider reading them all first and use a hash to store the results for comparison later, for example

#!perl
# test.pl
use strict;
use warnings;
use Cwd;
use Digest::MD5;
use Data::Dump 'pp';

my $sourcedirectory = cwd(); # '//path/to/directory'

# create test file with md5 value
open my $io_handle,'<','test.pl' or die "$!";
my $md5 = Digest::MD5->new;
$md5->addfile($io_handle);
close  $io_handle;
open OUT,'>','test.md5' or die "$!";
print OUT $md5->hexdigest;
close OUT;

# read md5 values into a hash
my $md5values = get_md5values($sourcedirectory);
pp $md5values;

sub get_md5values {
    
  my $sourcefiles = shift;
  opendir(DIR, $sourcefiles) 
    or die "Can't open directory, $!";
  my @md5_files = grep /\.md5$/,readdir(DIR);
  closedir(DIR);
  my %md5 = ();
  for my $filename (@md5_files){
    my $count = 0;
    open MD5,'<',$filename or die "$!";
    while (<MD5>){
      chomp;
      $md5{$filename} .= $_;
      ++$count;
    }
    close MD5;
    print "$count lines read from $filename\n";
  }
  return \%md5;   
}
[download]