07 - More Process Shceduling (T or F)

SJF chooses the process with the shortest next CPU burst

True

SRTF is the preemptive version of SJF

True

The OS can know the exact future CPU burst length of a process

False

SJF/SRTF relies on predicting CPU burst times

True

Exponential averaging is used to estimate the next CPU burst

True

p(n) represents the actual CPU burst time

False

t(n) represents the actual measured CPU burst time

True

p(1) is usually given as a constant or system average

True

The exponential averaging formula is p(n+1)=w·t(n)+(1-w)·p(n)

True

The value w must be between 0 and 1

True

If w=1, then p(n+1)=t(n)

True

If w=0, predictions never change from p(1)

True

Higher w means recent CPU bursts affect prediction more

True

Lower w means older history affects prediction more

True

Priority scheduling can cause starvation

True

Starvation means a process may wait forever

True

Aging helps reduce starvation

True

Aging increases the priority of waiting processes over time

True

Round-Robin gives each process a fixed time quantum

True

If the time quantum is very large, Round-Robin behaves like FCFS

True

If the time quantum is very small, context switching overhead increases

True

More context switching increases turnaround time

True

Rule of thumb: 80% of CPU bursts should be ≤ quantum

True

Multilevel Queue Scheduling assigns processes into fixed groups

True

In Multilevel Queue Scheduling, processes can move between queues freely

False

Multilevel Feedback Queue Scheduling allows processes to move between queues

True

In MFQS, CPU-heavy processes are moved to lower-priority queues

True

In MFQS, interactive/I/O-bound processes are moved to higher-priority queues

True

Multiprocessor scheduling is more complex than single-processor scheduling

True

A heterogeneous multiprocessor system has identical CPUs

False

A homogeneous multiprocessor system has identical CPUs

True

Using separate ready queues may leave one CPU idle while another is busy

True

A shared ready queue can help load balancing in multiprocessor systems

True

Self-scheduling is also called symmetric multiprocessing (SMP)

True

Master-slave scheduling is also called asymmetric multiprocessing (ASMP)

True

Hard real-time systems guarantee deadlines

True

Soft real-time systems guarantee deadlines

False

Hard real-time systems may reject a process if deadlines can’t be guaranteed

True

Deterministic modeling uses a fixed predetermined workload

True

Queuing models are math-intensive and use probability distributions

True

Simulation is usually more expensive than deterministic modeling

True

Implementation testing is the most expensive evaluation method

True