The unpack approach is probably fastest, but here's a regex way that doesn't require shortening the output array:

>perl -wMstrict -le
"my $len = 4;
 my $seq = 'qwertyuiop'
 ;;
 my @contigs = $seq =~ m{ (?= (.{$len})) }xmsg;
 printf qq{'$_' } for @contigs;
"
'qwer' 'wert' 'erty' 'rtyu' 'tyui' 'yuio' 'uiop'
[download]