This results have been generated on a NON-EMPTY system. This has virtually no
influence for most programs, as they rely on a clean CPU timer.

However, the results of the programs that use a wall-clock timer are (much)
worse than might be expected from an empty single-user system. Therefore,
the results for the programs MOD3A and MOD3B which measure I/O do not reflect
the actual performance. They are included only for completeness' sake.
