Well, I am currently working on Sort::Key::Radix that uses a Radix Sort as the sorting algorithm and has identical usage to Sort::Key.

When sorting large data sets and for some kinds of data, as for instance small numbers (as shown on the OP), it can be much faster than Sort::Key.

use Sort::Key::Radix qw(ukeysort);

my @by_order = ukeysort { $_->{order} } @data;
[download]

...though I am still getting some FAIL reports from the CPAN testers.