Well, if you don't mind a quick and dirty inefficient program, here's
something that I think does what you want. You cd into the top of the
directory tree you want to look at, for example "cd /home/abc".
Then run it "find_big_files.pl > /tmp/log_file". The results should be
dumped into the log file. Of course, you have to put find_big_files.pl
somewhere on your $PATH

It's very slow to output anything if you run it in a directory that
has many files/subdirectories. If you cd into / and run it, better
plan to leave it overnight, and I can't speak for how much memory it
might use.

It may have bugs, I wrote it about as fast as I could type. For the
same reason, it's not very efficient.

But it does seem to work.

Feel free to improve and/or share it, under any version of the GPL you prefer.

I'll bet that after I post this, someone will come up with a one-liner
solution. :-)

#!/usr/bin/perl

# find_big_files.pl
#
# abc 5/2008

use warnings;
use strict;

$|++;

my ($in_pipe);
my ($current_line);
my (%file_hash);
my ($key);
my ($entry);

unless (open ($in_pipe, "/usr/bin/find . -exec /bin/ls -ld {} \\; |"))
{
die "***ERROR*** Could not open pipe from find command, $!\n";
}

while ($current_line = <$in_pipe>)
{
chomp ($current_line);

$key = (split (/\s+/, $current_line))[4];


if ($key)
{
push @{$file_hash{$key}}, $current_line;
}
}

foreach $key (sort {$b <=> $a} keys %file_hash)
{
foreach $entry (@{$file_hash{$key}})
{
print "$entry\n";
}
}

close ($in_pipe);