1. Advanced Operating Systems - Spring 2009 Lecture 13 – February 23, 2009
Dan C. Marinescu
Office: HEC 439 B.
Office hours: M, Wd 3 – 4:30 PM.
TA: Chen Yu
Office: HEC 354.
Office hours: M, Wd – 3:00 PM.

2. Last, Current, Next Lecture
Last time:
Birth-death processes
M/M/1 systems
Today
CPU Scheduling
Scheduling Algorithms
M/M/m systems
Next time:
Caching and Virtual Memory

3. Histogram of CPU-burst Times

4. Dispatcher
Dispatcher module gives control of the CPU to the process selected by the short-term scheduler; this involves:
switching context
switching to user mode
jumping to the proper location in the user program to restart that program
Dispatch latency – time it takes for the dispatcher to stop one process and start another running

5. Performance metrics and objectives
CPU Utilization  Fraction of time CPU does useful work over total time
Throughput  Number of jobs finished per unit of time
Turnaround time  Time spent by a job in the system
Response time  Time to get the results
Waiting time  Time waiting to start processing
All these are random variables  we are interested in averages!!
The objectives of system managers (M) and users (U):
Maximize CPU utilization M
Maximize throughput  M
Minimize turnaround time  U
Minimize waiting time  U
Minimize response time  U

6. First-Come, First-Served (FCFS)
Process Burst Time
P1 24
P2 3
P3 3
Processes arrive in the order: P1  P2  P3 Gantt Chart for the schedule:
Waiting time for P1 = 0; P2 = 24; P3 = 27
Average waiting time: ( )/3 = 17
Convoy effect short process behind long process P1 P2 P3 24 27 30

7. FCFS Scheduling (Cont’d.)
Now processes arrive in the order: P2  P3  P1
Gantt chart:
Waiting time for P1 = 6; P2 = 0; P3 = 3
Average waiting time: ( )/3 = 3
Much better!! P1 P3 P2 6 3 30

8. Shortest-Job-First (SJF)
Use the length of the next CPU burst to schedule the process with the shortest time.
SJF is optimal minimum average waiting time for a given set of processes
Two schemes:
Non-preemptive  the process cannot be preempted until completes its CPU burst
Preemptive  if a new process arrives with CPU burst length less than remaining time of current executing process, preempt. known as Shortest-Remaining-Time-First (SRTF)

9. Non-preemptive SJF example
Process Arrival Time Burst Time P P P P
SJF (non-preemptive)
Average waiting time = ( )/4 = 4 P1 P3 P2 7 3 16 P4 8 12

10. Shortest-Remaining-Time-First (SRTF) (Preemptive SJF) example
Process Arrival Time Burst Time P P P P
Shortest-Remaining-Time-First
Average waiting time = ( )/4 = 3 P1 P3 P2 4 2 11 P4 5 7 16

11. Estimating the length of next CPU burst
Done using the length of previous CPU bursts, using exponential averaging

12. Exponential averaging
 =0
n+1 = n
Recent history does not count
 =1
n+1 =  tn
Only the actual last CPU burst counts
If we expand the formula, we get:
n+1 =  tn+(1 - ) tn -1 + …
+(1 -  )j  tn -j + …
+(1 -  )n +1 0
Since both  and (1 - ) are less than or equal to 1, each successive term has less weight than its predecessor

13. Predicting the length of the next CPU burst

14. Priority scheduling
Each process has a priority and the process with the highest priority (smallest integer  highest priority) is scheduled next.
Preemptive
Non-preemptive
SJF is a priority scheduling where priority is the predicted next CPU burst time
Problem  Starvation – low priority processes may never execute
Solution to sarvation  Aging – as time progresses increase the priority of the process

15. Round Robin (RR)
Each process gets a small unit of CPU time (time quantum), usually milliseconds. After this time has elapsed, the process is preempted and added to the end of the ready queue.
If there are n processes in the ready queue and the time quantum is q, then each process gets 1/n of the CPU time in chunks of at most q time units at once. No process waits more than (n-1)q time units.
Performance
q large  FIFO
q small  q must be large with respect to context switch, otherwise overhead is too high

16. RR with time slice q = 20 Process Burst Time P1 53 P2 17 P3 68 P4 24
Typically, higher average turnaround than SJF, but better response P1 P2 P3 P4 20 37 57 77 97
117
121
134
154
162

17. Time slice (quantum) and context switch time

18. Turnaround time function of time quantum

19. Multilevel queue
Ready queue is partitioned into separate queues each with its own scheduling algorithm :
foreground (interactive)  RR
background (batch)  FCFS
Scheduling between the queues
Fixed priority scheduling - (i.e., serve all from foreground then from background). Possibility of starvation.
Time slice – each queue gets a certain amount of CPU time which it can schedule amongst its processes; i.e.,
80% to foreground in RR
20% to background in FCFS

20. Multilevel Queue Scheduling

21. Multilevel feedback queue
A process can move between the various queues; aging can be implemented this way
Multilevel-feedback-queue scheduler characterized by:
number of queues
scheduling algorithms for each queue
strategy when to upgrade/demote a process
strategy to decide the queue a process will enter when it needs service

22. Multilevel feedback queue example
Three queues:
Q0 – RR with time quantum 8 milliseconds
Q1 – RR time quantum 16 milliseconds
Q2 – FCFS
Scheduling
A new job enters queue Q0 which is served FCFS. When it gains CPU, job receives 8 milliseconds. If it does not finish in 8 milliseconds, job is moved to queue Q1.
At Q1 job is again served FCFS and receives 16 additional milliseconds. If it still does not complete, it is preempted and moved to queue Q2.

23. Multilevel Feedback Queues

24. Unix scheduler
The higher the number quantifying the priority the lower the actual process priority.
Priority = (recent CPU usage)/2 + base
Recent CPU usage  how often the process has used the CPU since the last time priorities were calculated.
Does this strategy raises or lowers the priority of a CPU-bound processes?
Example:
base = 60
Recent CPU usage: P1 =40, P2 =18, P3 = 10

25. Multiple-processor scheduling
CPU scheduling more complex when multiple CPUs are available
Homogeneous processors within a multiprocessor
Load sharing
Asymmetric multiprocessing – only one processor accesses the system data structures, alleviating the need for data sharing

26. Comparison of scheduling algorithms
Round Robin
FIFO
MFQ
Multi-Level
Feedback Queue
SFJ
Shortest Job First
SRJN
Shortest Remaining Job Next
Throughput
Response
time
May be low is quantum is too small
Shortest average response
time if quantum chosen correctly
Not emphasized
May be poor
Good for I/O bound but poor for CPU-bound
processes
High
Good for short processes
But maybe poor for longer processes

27. Comparison of scheduling algorithms (cont’d)
Round Robin
FIFO
MFQ
Multi-Level
Feedback Queue
SFJ
Shortest Job First
SRJN
Shortest Remaining Job Next
IO-bound
Infinite postponement
No distinction
between
CPU-bound and
Does not occur
Gets a high priority if CPU-bound processes are present
May occur for CPU bound processes
May occur for processes with long estimated running times

28. Comparison of scheduling algorithms (cont’d)
Round Robin
FIFO
MFQ
Multi-Level
Feedback Queue
SFJ
Shortest Job First
SRJN
Shortest Remaining Job Next
Overhead
CPU-bound
Low
No distinction
between
CPU-bound and
IO-bound
The lowest
Can be high Complex data structures and processing routines
Gets a low priority if IO-bound processes are present
Can be high Routine to find to find the shortest job for each reschedule
Can be high Routine to find to find the minimum remaining time for each reschedule

29. Scheduling algorithms
A scheduling problems is defined by :
The machine environment
A set of side constrains and characteristics
The optimality criterion
Machine environments:
1  One-machine.
P  Parallel identical machines
Q  Parallel machines of different speeds
R  Parallel unrelated machines
O  Open shop. m specialized machines; a job requires a number of operations each demanding processing by a specific machine
F  Floor shop

30. One-machine environment
n jobs 1,2,….n.
pj amount of time required by job j.
rj  the release time of job j, the time when job j is available for processing.
wj  the weight of job j.
dj due time of job j; time job j should be completed.
A schedule S specifies for each job j which pj units of time are used to process the job.
CSj  the completion time of job j under schedule S.
The makespan of S is: CSmax = max CSj
The average completion time is

31. One-machine environment (cont’d)
Average weighted completion time:
Optimality criteria  minimize:
the makespan CSmax
the average completion time :
The average weighted completion time:
 the lateness of job j
 maximum lateness of any job under schedule S. Another optimality criteria, minimize maximum lateness.

32. Priority rules for one machine environment
Theorem: scheduling jobs according to SPT – shortest processing time is optimal for
Theorem: scheduling jobs in non-decreasing order of
is optimal for

33. Earliest deadline first (EDF)
Dynamic scheduling algorithm for real-time OS.
When a scheduling event occurs (task finishes, new task released, etc.) the priority queue will be searched for the process closest to its deadline. This process will then be scheduled for execution next.
EDF is an optimal scheduling preemptive algorithm for uniprocessors, in the following sense: if a collection of independent jobs, each characterized by an arrival time, an execution requirement, and a deadline, can be scheduled (by any algorithm) such that all the jobs complete by their deadlines, the EDF will schedule this collection of jobs such that they all complete by their deadlines.

34. EDF
The schedulability test for EDF is: In this case U = 1/8 +2/5 + 4/10 = = 92.5% It has been proved that the problem of deciding if it is possible to schedule a set of periodic processes is NP-hard if the periodic processes use semaphores to enforce mutual exclusion.
Process
Execution Time
Period P1 1 8 P2 2 5 P3 4 10

35. Priority Inversion
A high priority process is blocked by a lower priority one.
Example: J1 and J3 share a data structure guarded by a binary semaphore S.
prty(J1) > prty(J2) > prty(J3).
J1 in initiated while J3 is in its critical section
When J1 attempts to enter the critical section it is blocked.
The duration of this blocking cannot be determined as because J3 can be preempted by a higher priority job J2. prty