Lecture 5: CPU Scheduling

Question 1

define CPU scheduling

Answer 1

select processes from ready state, and allocate the CPU to them

Question 2

nonpreemptive scheduling

Answer 2

take place when a process:
- switch from running to waiting state
- terminate

Question 3

preemptive scheduling

Answer 3

take place when a process:
- switch from running to ready state
- switch from waiting to ready state

Question 4

nonpreemptive vs preemptive scheduling

Answer 4

• Preemptive scheduling allows a running process to be interrupted by a high priority process
• in non-preemptive scheduling, any new process has to wait until the running process finishes its CPU cycle.

Question 5

3 tasks involve in dispatcher module

Answer 5

• switching context
• switching to user mode
• jumping to proper location to restart program

Question 6

define dispatch latency

Answer 6

time taken for dispatcher to stop a process and start a new running process

Question 7

5 scheduling criteria

Answer 7

• CPU utilization: keep CPU as busy as possible
• throughput: number of process completed per time unit
• turnaround time: amount of time to complete a process
• waiting time: amount of time of a process waiting in ready queue
• response time: time taken from the request submitted until the first response (not output)

Question 8

convoy effect

Answer 8

short process behind long process, causing many short process get blocked
- one CPU bound and many I/O bound process

Question 9

ready queue is separated into 2 queues

Answer 9

• background (batch) using FCFS
• foreground (interactive) using RR

Question 10

problem of priority scheduling

Answer 10

starvation
- process with long burst time will never execute

Question 11

solution of starvation

Answer 11

aging
- priority of process increase by the time progress

Question 12

formula for AWT (average waiting time)

Answer 12

turnaround time = finished time - arrived time
waiting time = turnaround time - burst time

Question 13

multilevel feedback queue

Answer 13

process can move between various queue

Question 14

parameters in multilevel feedback queue

Answer 14

• number of queue
• scheduling algorithm for each queue
• method determined when to upgrade a process
• method determined when to demote a process
• method determined which queue the process will enter

Question 15

2 processor affinity

Answer 15

soft affinity - process has policy to attempt to keep running in the same processor
hard affinity - process can specify the subset of processor to run in

Question 16

multiprocessor core

Answer 16

multiple processor core in one physical chip

Question 17

benefits of multiprocessor core

Answer 17

• faster
• less power consumed

Question 18

Little’s law

Answer 18

in steady state, processes leaving queue must equal to arriving queue

• n = l x W
  where,
  n = average queue length
  l = average waiting time in queue
  W = average arrival time in queue

Question 19

differentiate FCFS, RR and Multilevel Feedback Queue

Answer 19

First Come First Serve Scheduling
- no discriminate
- no matter short or long process
- based on arrival time

Round Robin Scheduling
- short process finish early
- shorter waiting time

Multilevel Feedback Queue
- try to complete short process first
- long process wait at queue, short process finish early

Question 20

benefits of RR

Answer 20

implement time quantum (q) in milliseconds
- no starvation
- better response (but higher turnaround than SJF)
- suitable for time-sharing system

Question 21

all scheduling method

Answer 21

• first come first serve (FCFS)
• preemptive SJF
• non-preemptive SJF
• priority scheduling
• round robin (RR)