Download P 3

‫بسم هللا الرحمن الرحيم‬
CPCS361 – Operating Systems I
Chapter 5: CPU Scheduling
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
Chapter 5: CPU Scheduling







Basic Concepts
Scheduling Criteria
Scheduling Algorithms
Thread Scheduling
Multiple-Processor Scheduling
Operating Systems Examples
Algorithm Evaluation
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
2
Objectives



To introduce CPU scheduling, which is the basis for
multiprogrammed operating systems
To describe various CPU-scheduling algorithms
To discuss evaluation criteria for selecting a CPUscheduling algorithm for a particular system
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
3
5.1 Basic Concepts



Maximum CPU utilization obtained with
multiprogramming
The objective of multiprogramming is to have
some process running at all times
The idea is to execute a process until it must wait,
typically for the completion of some I/O request.
In a simple computer system, the CPU just sits
idle. All this waiting time is wasted; no useful work
is accomplished
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
4
Basic Concepts




With multiprogramming, we try to use this time
productively
Several processes are kept in memory at one time;
when one process has to wait, the OS takes the
CPU away from that process and gives the CPU to
another process. This pattern continues
Scheduling of this kind is a fundamental OS
function
Scheduling is central to OS design
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
5
Basic Concepts



The success of CPU scheduling depends on an observed
property of processes
CPU–I/O Burst Cycle – Process execution consists of a
cycle of CPU execution and I/O wait
Processes alternate between these two states: CPU burst
 I/O burst  CPU burst  I/O burst  … CPU burst
(system request to terminate execution)
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
6
Basic Concepts


The durations of CPU bursts have been measured
extensively. Although they vary greatly from process to
process and from computer to computer, they tend to have
a frequency curve similar to that shown in the CPU burst
distribution (histogram)
The curve is generally characterized as exponential, with a
large number of short CPU bursts and a small number of
long CPU bursts



I/O-bound program: many short CPU bursts
CPU-bound program: few long CPU bursts
Distribution is important in the selection of an appropriate CPUscheduling algorithm
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
7
Alternating Sequence of CPU And I/O Bursts
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
8
Histogram of CPU-burst Times
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
9
CPU Scheduler

Selects from among the processes in memory (ready
queue) that are ready to execute, and allocates the CPU
to one of them

Selection is carried by the short-term scheduler (CPU
scheduler)
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
10
CPU Scheduler

CPU scheduling decisions may take place when a process:
1. Switches from running to waiting state (e.g. I/O request)
2. Switches from running to ready state (e.g. Interrupt)
3. Switches from waiting to ready (e.g. I/O completion)
4. Terminates
P
NP
Running
Ready
NP
Terminated
Waiting
P
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
11
CPU Scheduler

Scheduling under 1 and 4 is nonpreemptive
(cooperative)

Once the CPU has been allocated to a process, the process
keeps the CPU until it releases the CPU either by terminating
or by switching to the waiting state


e. g. Windows 3.x, Windows 95
The only method that can be used on certain hardware
platforms, because it does not require special hardware (e.g. a
timer) needed for preemptive scheduling
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
12
CPU Scheduler

Scheduling under 2 and 3 is preemptive


Unfortunately, it incurs a cost associated with access to shared
data
Case of two processes that share data; while one is updating
the data, it is preempted so that the second process can run.
The second process then tries to read the data, which are in an
inconsistent state. A new mechanism to coordinate access to
shared data is needed
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
13
CPU Scheduler

Scheduling under 2 and 3 is preemptive

Preemptive also affects the design of OS kernel


During the processing of a system call, the kernel may be busy with an
activity on behalf of a process. Such activity may involve changing
important kernel data (for instance, I/O queues). What happens if the
process is preempted in the middle of these changes and the kernel
(or device driver) needs to read or modify the same structure? Chaos
ensues
Certain OSs (including UNIX) deal with this problem by waiting
either for a system call to complete or for an I/O block to take place
before doing a context switch
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
14
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
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
15
5.2 Scheduling Criteria



Different CPU scheduling algorithms have different
properties
Many criteria have been suggested for comparing
scheduling algorithms
Which characteristics are used for comparing CPU
scheduling algorithms can make substantial difference
in which algorithm is judged to be best.
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
16
5.2 Scheduling Criteria

The criteria include the following

CPU utilization – keep the CPU as busy as possible




Conceptually from 0 to 100%
In real system it should range from 40% (lightly used system) to 90%
(heavily used system)
Throughput – # of processes that complete their execution per
time unit
Turnaround time – amount of time to execute a particular
process


Interval from the time of submission to the time of completion of a
process
Sum of the periods spent waiting to get into memory, waiting in the ready
queue, executing on the CPU, and doing I/O
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
17
5.2 Scheduling Criteria

The criteria include the following


Waiting time – amount of time a process has been waiting in the
ready queue (sum of the periods spent waiting in the ready queue)
Response time – amount of time it takes from when a request
was submitted until the first response is produced, not output (for
time-sharing environment)
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
18
Scheduling Algorithm Optimization Criteria
 It is desirable to



Under some circumstances, it is desirable to optimize the
minimum or maximum values rather than the average


Maximize: CPU utilization, throughput
Minimize: turnaround time, waiting time, response time
e. g. to guarantee all users get good service, we may want to
minimize the maximum response time
A system with reasonable and predictable response time
may be considered more desirable than a system that is
faster on the average but is highly variable
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
19
5.3 Scheduling Algorithms


CPU scheduling deals with the problem of deciding which
of the processes in the ready queue is to be allocated to
the CPU
There are many different CPU scheduling algorithms:






First Come First Serve (FCFS)
Shortest Job First (SJF)
Priority
Round-Robin (RR)
Multilevel Queue (MLQ)
Multilevel Feedback-Queue (MLFQ)
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
20
First-Come, First-Served (FCFS) Scheduling

Process Burst Time
P1
24
P2
3
P3
3
Suppose that the processes arrive in the order: P1, P2, P3
The Gantt Chart for the schedule is:
P1
0


P3
P2
24
27
30
Waiting time for P1 = 0; P2 = 24; P3 = 27
Average waiting time: (0 + 24 + 27)/3 = 17
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
21
FCFS Scheduling (Cont)
Suppose that the processes arrive in the order
P2 , P3 , P1
 The Gantt chart for the schedule is:
P2
0




P3
3
P1
6
30
Waiting time for P1 = 6;P2 = 0; P3 = 3
Average waiting time: (6 + 0 + 3)/3 = 3
Much better than previous case
Convoy effect short process behind long process
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
22
FCFS Scheduling (Cont)



The average waiting time under FCFS policy is generally
not minimal and may vary substantially if the process’s
CPU burst times vary greatly
FCFS scheduling algorithm is nonpreemptive
FCFS is particularly troublesome for time-sharing systems,
where it is important that each user get a share of the
CPU at regular intervals. It would be disastrous to allow
one process to keep the CPU for an extended period
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
23
Shortest-Job-First (SJF) Scheduling



Associate with each process the length of its next CPU
burst. Use these lengths to schedule the process with
the shortest time
If the next CPU bursts of two processes are the same,
FCFS is used to break the tie
SJF is optimal – gives minimum average waiting time for a
given set of processes
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
24
Example of SJF

Process Arrival Time
P1
0.0
P2
2.0
P3
4.0
P4
5.0
SJF scheduling chart
P4
0


P1
3
Burst Time
6
8
7
3
P3
9
P2
16
24
Average waiting time = (3 + 16 + 9 + 0) / 4 = 7
With FCFS average waiting time = 10.25
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
25
Determining Length of Next CPU Burst




The real difficulty with the SJF algorithm is knowing the
length of the next CPU burst
For long-term scheduling we can use as the length the
process time limit that a user specifies when he submits
the job
SFJ is used frequently in long-term scheduling
Although SFJ algorithm is optimal, it cannot be
implemented at the level of short-term scheduling. There
is no way to know the length of the next CPU burst
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
26
Determining Length of Next CPU Burst



One approach is to try to estimate SFJ scheduling. We
may not know the length of the next CPU burst, but we
may be able to predict its value
Expect that the next CPU burst will be similar in length
to the previous ones
Thus, by computing an approximation of the length of the
next CPU burst, we can pick the process with the
shortest predicted CPU burst
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
27
Determining Length of Next CPU Burst


Can only estimate the length
Can be done by using the length of previous CPU bursts,
using exponential averaging
1. t n  actual length of n th CPU burst
2.  n 1  predicted value for the next CPU burst
3.  , 0    1


4. Define :  nn+1
1   t n  1    n .

Example: Assume  = ½ , 0 = 10, and t0 = 6
Then, 1 = ½ (6) + ½ (10) = 8, t1= 4
2 = ½ (4) + ½ (8) = 6
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
28
Determining Length of Next CPU Burst
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
29
Examples of Exponential Averaging

=0



=1



n+1 = n
Recent history does not count
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
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
30
Priority Scheduling



A priority number (integer) is associated with each
process
SJF is a special case of priority scheduling
The CPU is allocated to the process with the highest
priority (smallest integer  highest priority)




Preemptive
Nonpreemptive
Equal-priority processes are scheduled in FCFS order
SJF is a priority scheduling where priority is the inverse
of the predicted next CPU burst time (larger CPU-burst
 lower the priority)
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
31
Priority Scheduling

Priority can be defined

Internally: Use some measurable quantity or quantities to
compute the priority of a process; e. g.





Time limits
Memory requirements
# open files
Ratio of average I/O burst to average CPU burst
Externally: set by criteria outside the OS; such as,



Importance of the process
Type and amount of funds being paid for computer use
Department sponsoring the work
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
32
Priority Scheduling

Priority scheduling: when a process arrives at the ready
queue, its priority is compared to the priority of the
currently running process


Preemptive: Preempt the CPU if the priority of the newly
arrived process is higher than the priority of the currently
running process
Nonpreemptive: Put the new process at the head of the ready
queue
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
33
Priority Scheduling


Problem  Starvation (Indefinite blocking)– low
priority processes may never execute in a heavily loaded
computer system
Solution  Aging – a technique if gradually increase the
priority of processes that wait in the system for a long
time; that is as time progresses increase the priority of
the process.


For example, increase the priority of a waiting process by 1
every 15 minutes
It would take no more than 32 hours for a pririty-127 process
to age to a priority-0 process
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
34
Priority Scheduling

Process
P1
P2
P3
P4
P5
The Gantt chart is:
P2
0

P5
1
Burst Time
10
1
2
1
5
Priority
3
1
4
5
2
P1
6
P3
16
P4
18
19
Average waiting time = 8.2 ms
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
35
Round Robin (RR)




Designed for time-sharing systems
Similar to FCFS scheduling, but preemption is added
to switch between processes
Each process gets a small unit of CPU time (time
quantum / time slice), usually 10-100 milliseconds.
After this time has elapsed, the process is preempted
and added to the end of the ready queue.
CPU scheduler goes around the ready queue,
allocating the CPU to each process for a time
interval up to 1 time quantum
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
36
Round Robin (RR)

CPU scheduler picks the first process from the ready
queue, sets a timer to interrupt after 1 time quantum,
and dispatches the process; two cases:


CPU burst  1 time quantum  process itself will
release the CPU voluntarily. Scheduler will then proceed
to the next process in the ready queue
CPU burst >1 time quantum  timer will go off and will
cause an interrupt, context switch will be executed, and
the process will be put at the tail of the ready queue,
scheduler will then select the next process in the ready
queue
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
37
Example of RR with Time Quantum = 4

Process
Burst Time
P1
P2
P3
24
3
3
The Gantt chart is:
P1
0


P2
4
P3
7
P1
10
P1
14
P1
18 22
P1
26
P1
30
Average waiting time = 17 /3 = 5.66 ms
Typically, higher average turnaround than SJF, but
better response
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
38
Round Robin (RR)


No process is allocated the CPU for more than 1 time
quantum unless it is the only runnable process
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.

Example: n = 5, q = 20 ms
Then, each process will get up to 20 ms every 100 ms
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
39
Round Robin (RR)

Performance: depends heavily on the size of the time
quantum



q large  FIFO (FCFS) policy
q small  q must be large with respect to context switch,
otherwise overhead is too high;
Example: CPU burst of p = 10 time units
If q = 12 time units, the process finishes in < 1 quantum, with no overhead
if q = 6, the process requires 2 quanta, resulting in context switch
if q = 1, then 9 context switch will occur, slowing execution of CPU
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
40
Time Quantum and Context Switch Time
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
41
Round Robin (RR)



Time quantum to be large with respect to the
context switch time
If the time quantum is about 10% of the quantum
time, then about 10% of the CPU time will be spent
in context switching
In practice, modern systems have time quantum
ranging from 10 to 100 milliseconds. Time required
for context switch is typically < 10ms (a small
fraction of the time quantum)
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
42
Round Robin (RR)


Turnaround time also depends on the size of the time
quantum
Average turnaround time of a set of processes does
not necessarily improve as the quantum time
increases

Example: 3 processes each 10 time units,
q = 1  average turnaround time = 29
q = 10  average turnaround time = 20
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
43
Round Robin (RR)


Average turnaround time can be improved if most
processes finish their next CPU burst in a single time
quantum
Although the time quantum should be large compared
with the context switch time, it should be not too
large


If the time quantum is too large, RR scheduling
degenerates to FCFS policy
Rule of thumb: 80% of the CPU bursts should be shorter
than time quantum
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
44
Turnaround Time Varies With The Time
Quantum
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
45
Multilevel Queue (MLQ)


Processes are classified into different groups
e.g. ready queue is partitioned into separate queues:





foreground (interactive)
background (batch)
They have different response-time requirements and so may
have different scheduling needs
FG processes may have priority (externally defined) over BG
processes
MLQ scheduling partitions the ready queue into several
separate queues

Processes are permanently assigned to one queue, generally
based on some property of the process, such as memory size,
process priority, or process type
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
46
Multilevel Queue (MLQ)

Each queue has its own scheduling algorithm



foreground – RR
background – FCFS
Scheduling must be done between the queues


Fixed priority preemptive scheduling; (i.e., serve all from
foreground then from background). Possibility of starvation. If
an interactive editing process entered the ready queue while a
batch process was running, the batch process would be
preempted.
Time slice – each queue gets a certain amount of CPU time
which it can schedule amongst its processes


80% to foreground queue in RR
20% to background queue in FCFS
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
47
Multilevel Queue Scheduling
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
48
Multilevel Feedback Queue (MLFQ)

In MLQ a process is permanently assigned to a queue
when it enter the system

Processes do not move from one queue to another



Advantage: low scheduling overhead
Disadvantage: inflexible
In MLFQ a process can move between the various
queues; aging can be implemented this way

The idea is to separate processes according to the
characteristics of their CPU bursts


A process that uses too much CPU time will be moved to lower –
priority queue. This scheme leaves I/O-bound and interactive
processes in the higher-priority queues
A process that waits too long in lower-priority queue may be moved
to a higher-priority queue
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
49
Multilevel Feedback Queue (MLFQ)


A process can move between the various queues;
aging can be implemented this way
Multilevel-feedback-queue scheduler defined by the
following parameters:





number of queues
scheduling algorithms for each queue
method used to determine when to upgrade a process
method used to determine when to demote a process
method used to determine which queue a process will
enter when that process needs service
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
50
Example of Multilevel Feedback Queue

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.
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
51
Multilevel Feedback Queues
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
52
5.4 Multiple-Processor Scheduling





With multiple CPUs load sharing becomes possible
CPU scheduling more complex when multiple CPUs are
available
Homogeneous (identical) processors within a
multiprocessor
Asymmetric multiprocessing – only one processor
accesses the system data structures and the other
processors execute only user code, reducing the need for
data sharing.
Symmetric multiprocessing (SMP) – each processor
is self-scheduling, all processes in common ready queue, or
each has its own private queue of ready processes
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
53
Multiple-Processor Scheduling

Processor Affinity



Data most recently accessed by the process populates the
cache for the processor; and as a result, successive
memory accesses by the process are often satisfied in
cache memory
If process migrates to another processor, the contents of
the cache memory must be invalidated for the processor
being migrated from, and the cache memory for the
processor being migrated to must be re-populated
Because of the high cost of invalidating and re-populating
caches, most SMP systems try to avoid migration of
processes from one processor to another and instead
attempt to keep a process running on the same processor.
This is known as: Processor affinity
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
54
Multiple-Processor Scheduling

Processor Affinity



Process has affinity for processor on which it is currently
running
Soft affinity: OS has a policy of attempting to keep a
process running on the same processor – but not
guaranteeing that it will do so
Hard affinity: Systems that provide system calls that support
allowing a process to specify that it is not to migrate to
another processors; e.g. Linux
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
55
Multiple-Processor Scheduling

Load Balancing



On SMP systems, it is important to keep the workload
balanced among all processors to fully utilize the benefits of
having more than one processor. Otherwise, one or more
processor may sit idle while other processors have high
workload along with lists of processes awaiting the CPU
Attempts to keep the workload evenly distributed across all
processors in an SMP system
Not needed on systems with common run queue
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
56
Multiple-Processor Scheduling

Load Balancing

Two approaches:



Push migration: A specific task periodically checks the load on each
processor, if it finds an imbalance, evenly distributed the load by moving
(or pushing) processes from overloaded to idle or less-busy
processors
Pull migration: Occurs when an idle processor pulls a waiting task
from a busy processor
Load balancing often counteracts the benefits of processor
affinity


No absolute rule concerning what policy is best
In some systems, an idle processor always pulls a process from a nonidle processor; and in other systems, processors are moved only if the
imbalance exceeds a certain threshold
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
57
5.7 Algorithm Evaluation


Deterministic modeling: takes a particular predetermined
workload and defines the performance of each algorithm
for that workload
Queuing models: Processes that are run vary from day to
day, so there is no static set of processes to use for
deterministic modeling

Solution: Use CPU and I/O-bursts distributions, and arrivaltime distribution to compute: throughput, utilization, waiting
time, and so on for most algorithms
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
58
5.7 Algorithm Evaluation

Simulations: Running simulations involves programming a
model of the computer system


The simulator modifies the system status to reflect the
activities of the devices, the processes, and the scheduler
As the simulation executes, statistics that indicate algorithm
performance are gathered and printed
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
59
Evaluation of CPU schedulers by Simulation
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
60
End of Chapter 5
Adopted from and based on Textbook: Operating System Concepts – 8th Edition, by Silberschatz, Galvin and Gagne
Updated and Modified by Dr. Abdullah Basuhail, CSD, FCIT, KAU, 1431H
61