AN UNFAIR SEMI-GREEDY REAL-TIME MULTIPROCESSOR SCHEDULING ALGORITHM.

Optimal real-time multiprocessor scheduling algorithms always achieve higher
processor utilization that is equal to the number of processors in the system. However,
optimality always comes at the expense of scheduling overheads in terms of task
preemptions and migrations which highly affect the practicability of the algorithm.
This is because most of these algorithms achieve optimality by adhering to the
fairness rule, in which tasks are forced to make progress in their executions at each
time quanta or at the end of each time slice in a fluid schedule model; which
corresponds to the deadlines of all tasks in the system. These preemptions and
migrations add extra overheads that must be added to the worst case execution
requirements of a task.