Only 5 million records because I have an image analysis that has been running for 2 days now using over half my memory and I don't want to interrupt it.

I retrieved the first 1000 records from the hash -- as near random as makes no difference -- because 10 wasn't representative. Retrieval time is 0.4 millisecs per.

The code:

#! perl -slw
use strict;
use Time::HiRes qw[ time ];

our $N //= 1000;

my $start = time;
my %dna; keys %dna = 5e6;
$dna{ <> } = <> until eof();
printf "Loading %d records took %d seconds\n",
    scalar keys %dna, time() - $start;

$start = time();
my( $id, $seq ) = each %dna for 1 .. $N;

printf "Took %.6f seconds per retrieval\n",
    ( time() - $start ) / $N;

__END__
[22:14:18.06] C:\test>junk48 -N=1e3 junk.dat
Loading 5000000 records took 37 seconds
Took 0.000414 seconds per retrieval
[download]