Chapter 5:
Process Scheduling
Outline
 Basic Concepts
 CPU-I/O Burst Cycle
 CPU Scheduler
 Preemptive Scheduling
 Dispatcher
 Scheduling Criteria
 Scheduling Algorithms
 First-Come, First-Serve Scheduling
 Shortest-Job-First Scheduling
 Priority Scheduling
 Round-Robin Scheduling
 Multilevel Queue Scheduling
 Multilevel Feedback-Queue Scheduling
Basic Concepts
 CPU: One of the primary computer resource
 Maximum CPU utilization obtained with
multiprogramming
Several processes are kept in memory
 CPU–I/O Burst Cycle – Process execution
consists of a cycle of CPU execution and I/O
wait.
 CPU burst distribution determines how to
schedule the processes.
Alternating Sequence of CPU And I/O
Bursts
Histogram of CPU-burst Times
Normally a process starts with frequent
short CPU bursts and then it stabilizes
and shifts to long but infrequent
CPU bursts
CPU-bound program
 I/O bound
processes usually
have many small
CPU bursts
 CPU bound
programs usually
have few but long
CPU bursts
CPU Scheduler
 Selects from among the processes in memory
that are ready to execute, and allocates the CPU
to one of them.
Is it the same as Short-term scheduler? … Yes
 CPU scheduling decisions may take place when
a process:
1. Switches from running to waiting state.
2. Switches from running to ready state.
3. Switches from waiting to ready.
4. Terminates.(Own will)
 Scheduling Scheme given at 1 and 4 is
nonpreemptive (Cooperative).
 All other scheduling is preemptive.
Preemptive vs Non Preemptive
Scheduling
 Non Preemptive scheduling
If a process is being executed then it keeps the CPU
until one or four conditions of previous slide occur.
MS Windows 3.x which was sort of multitasking system
used this method
Not at all efficient.
 Preemptive
A process in running state can be brought back to ready
state without any of the conditions of previous slide
being met.
Can you guess how it is done? 
Problems with preemptive
Although better than non-preemptive still
suffers from synchronization problems
E.g.
Process P and Q share some data (can be
memory location or a file)
P does some calculations changes some of the
data but before it finishes its time expires
Q reads the data but is the data valid?
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.
Should be minimum.
Scheduling Criteria
Following criteria is kept under consideration while designing a scheduler
 CPU utilization – Keep the CPU as busy as possible
 Throughput – # of processes that complete their execution
per unit time (Given Time).
 Turnaround time – Amount of time to execute a particular
process (Sum of ready, waiting, running, loading times,I/O)
 Waiting time – Amount of time a process has been 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 interactive environment)
 Time it takes to start responding
 Not the time it takes to output the response
Optimization Criteria
So the best CPU Scheduling algorithm is
which …
Maximizes CPU utilization
Maximizes throughput
Minimizes turnaround time
Minimizes waiting time
Minimizes response time
Outline
 Basic Concepts
 CPU-I/O Burst Cycle
 CPU Scheduler
 Preemptive Scheduling
 Dispatcher
 Scheduling Criteria
 Scheduling Algorithms
 First-Come, First-Serve Scheduling
 Shortest-Job-First Scheduling
 Priority Scheduling
 Round-Robin Scheduling
 Multilevel Queue Scheduling
 Multilevel Feedback-Queue Scheduling
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
P2
24
P3
27
 Waiting time for P1 = 0; P2 = 24; P3 = 27
 Average waiting time: (0 + 24 + 27)/3 = 17
30
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.
CPU bound processes will get and hold the CPU
Convoy effect short process behind long process (Bad
CPU utilization)
Shortest-Job-First (SJR) Scheduling
 Associate with each process the length of its
next CPU burst. Use these lengths to schedule
the process with the shortest time.
 Two schemes:
 nonpreemptive – once CPU given to the process it cannot be
preempted until it completes its CPU burst.
 preemptive – if a new process arrives with CPU burst length less
than remaining time of current executing process, preempt. This
scheme is know as the
Shortest-Remaining-Time-First (SRTF).
 SJF is optimal – gives minimum average waiting
time for a given set of processes.
 Shortest-next-cpu-burst algorithm
Example of Non-Preemptive SJF
Process
P1
P2
P3
P4
Burst Time
6
8
7
3
 SJF
8
P1
P4
0
3
P3
9
P2
16
 Average waiting time = (3+16+9+0)/4=7
24
FCFS=10.25
Shortest-Job-First (SJR) Scheduling
 Good for long-term Scheduler in batch systems
User specifies process time during job submission
Lower value may mean faster response
Too low a value will cause a time-limit-exceeded error
 Difficult to implement at the level of short-term
CPU scheduler
 Determining Length of Next CPU Burst
Can only predict/ estimate the length of next CPU burst.
Can be done by using the length of previous CPU bursts,
using exponential averaging.
Examples of Exponential Averaging
  =0
n+1 = n
Recent history does not count.
  =1
n+1 =αtn +(1- α) n
tn: length of the nth CPU burst
n+1 : Our predicted Value
α: weight of recent & past
history
0<= α<=1
 n+1 = tn
Only the actual last CPU burst counts.
 Since both  and (1 - ) are less than or equal to 1, each
successive term has less weight than its predecessor.
Prediction of the Length of the Next CPU Burst
Initial guess for T(1) = 10 ms
Actual 6
Determination of 2 Cycle based on actual (6) and last guess (10)
T(2) = 0.5 x 6 + 0.5 x 10 = 8 ms
Determination of 3 cycle based on actual(4) and last guess (8)
T(3) = 0.5 x 4 + 0.5 x 8 = 6 ms
… Now you can do it 
 For example: Suppose a process p is given a default expected burst length
of 5 time units. When it is run, the actually burst lengths are 10,10,10,1,1,1
(although this information is not known in advance to any algorithm). The
prediction of burst times for this process works as follows.
Let e(1) = 5, as a default value.
When process p runs, its first burst actually runs 10 time units, so,
a(1) = 10.
e(2) =0.5 * e(1) * 0.5 + a(1) = 0.5 * 5 + 0.5 * 10 = 7.5
This is the prediction for the second cpu burst
The actual second cpu burst is 10. So the prediction for the third cpu burst is:
e(3) = 0.5 * e(2) * 0.5 + a(2) = 0.5 * 7.5 + 0.5 * 10 = 8.75
e(4) = 0.5 * e(3) * 0.5 + a(3) = 0.5 * 8.75 + 0.5 * 10 = 9.38,
So, we predict that the next burst will be close to 10 (9.38) because the
recent bursts have been of length 10. At this point, it happens that the
process starts having shorter bursts, with a(4) = 1, so the algorithm
gradually adjusts its estimated cpu burst (prediction)
e(5) = 0.5 * e(4) * 0.5 + a(4) = 0.5 * 9.38 + 0.5 * 1 = 5.19
e(6) = 0.5 * e(5) * 0.5 + a(5) = 0.5 * 5.19 + 0.5 * 1 = 3.10
e(7) = 0.5 * e(6) * 0.5 + a(6) = 0.5 * 3.10 + 0.5 * 1 = 2.05
Example of Preemptive SJF
(Shortest Remaining Time first)
Process Arrival Time
P1
0
P2
1
P3
2
P4
3
 SJF (preemptive)
P1
0 1
P2
P1
P4
5
Burst Time
8
4
9
5
10
P3
17
 Average waiting time = 6.5
 Non-Preemptive Average waiting time = 6.5
26
Outline
 Basic Concepts
 CPU-I/O Burst Cycle
 CPU Scheduler
 Preemptive Scheduling
 Dispatcher
 Scheduling Criteria
 Scheduling Algorithms
 First-Come, First-Serve Scheduling
 Shortest-Job-First Scheduling
 Priority Scheduling
 Round-Robin Scheduling
 Multilevel Queue Scheduling
 Multilevel Feedback-Queue Scheduling
Priority Scheduling
 A priority number (integer) is associated with each
process
 The CPU is allocated to the process with the highest
priority (smallest integer  highest priority).
 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  Aging – as time progresses increase the
priority of the process.
Round Robin (RR)
 Each process gets a small unit of CPU time (time
quantum), usually 10-100 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.
Example of RR with Time Quantum
= 20
Process
P1
P2
P3
P4
 The Gantt chart is:
P1
0
P2
20
37
P3
P4
57
P1
77
Burst Time
53
17
68
24
P3
97 117
P4
P1
P3
P3
121 134 154 162
 Typically, higher average turnaround than SJF,
but better response.
Time Quantum and Context Switch Time
 If context switching time is approx 10 percent of the time quantum,
then about 10 percent of the CPU time will be spent in context
switching
 Most modern systems have time quanta ranging from 10 to 100
milliseconds
 Context switch typically takes less than 10 microseconds
Remember: Context switch takes time 
Turnaround Time Varies With The Time
Quantum
80% of CPU
bursts should be
shorter than the
time quantum
Average
Turnaround time doesn’t necessarily improve as the time-quantum size increases)
Outline
 Basic Concepts
 CPU-I/O Burst Cycle
 CPU Scheduler
 Preemptive Scheduling
 Dispatcher
 Scheduling Criteria
 Scheduling Algorithms
 First-Come, First-Serve Scheduling
 Shortest-Job-First Scheduling
 Priority Scheduling
 Round-Robin Scheduling
 Multilevel Queue Scheduling
 Multilevel Feedback-Queue Scheduling
Multilevel Queue
 Ready queue is partitioned into separate queues:
foreground (interactive)
background (batch)
 Each queue has its own scheduling algorithm,
foreground – RR
background – FCFS
 Scheduling must be done 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
Multilevel Queue Scheduling
Multilevel Feedback Queue
 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
Example of Multilevel Feedback
Queue
 Consider a case if we have three queues:
Q0 – time quantum 8 milliseconds
Q1 – 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 tail of
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 tail of queue Q2.
Multilevel Feedback Queues
Outline
 Basic Concepts
 CPU-I/O Burst Cycle
 CPU Scheduler
 Preemptive Scheduling
 Dispatcher
 Scheduling Criteria
 Scheduling Algorithms
 First-Come, First-Serve Scheduling
 Shortest-Job-First Scheduling
 Priority Scheduling
 Round-Robin Scheduling
 Multilevel Queue Scheduling
 Multilevel Feedback-Queue Scheduling
 Multi-Processor Scheduling
Multiple-Processor Scheduling
 CPU scheduling more complex when multiple
CPUs are available.
 Consider Homogeneous processors within a
multiprocessor system.
 Load sharing
 Asymmetric multiprocessing – only one
processor accesses the system data structures,
alleviating the need for data sharing.
 SMP: Each processor is self scheduling
May have separate ready queues
Multiple-Processor Scheduling –Process Affinity
 Cache memory?
The Data most recently accessed populates it
 Process migration to another process
Invalidation & Repopulation
 Some OS attempt to keep a process running on
the same processor –Affinity
 Soft Affinity: Attempt to keep a process running
on the same processor but not guaranteeing
 Hard Affinity: Process cannot migrate to another
processor
Multiple-Processor Scheduling –Load Balancing
 Utilize the full benefits of multi-processors
 Load balancing: Distribute the workload evenly across all
processors.
 Necessary where each processors has its own private
queue of eligible processes
 Not necessary where processors have common run
queue
 Modern OS --private queue or common?
 Push Migration: move processes from overloaded to idle
processors
 Pull Migration: Idle processor pulls a waiting process
from a busy processor
 Any link with Process Affinity?
Multiple-Processor Scheduling
Symmetric Multithreading
 Multiple logical rather than physical processors
to run several threads concurrently (HPT)
 Create logical processors on the same physical
processor—a view of many Processors to OS
Each has its own machine-state registers
Own interrupt handling
Logical
CPU
Logical
CPU
Logical
CPU
Logical
CPU
Physical CPU
Physical CPU
System Bus
Multiple-Processor Scheduling
Symmetric Multithreading
SMT: a feature provided by hardware
Os needn’t be designed differently
Performance gains –If OS is aware.
But why?
**Scheduler**
Outline
 Basic Concepts
 CPU-I/O Burst Cycle
 CPU Scheduler
 Preemptive Scheduling
 Dispatcher
 Scheduling Criteria
 Scheduling Algorithms






First-Come, First-Serve Scheduling
Shortest-Job-First Scheduling
Priority Scheduling
Round-Robin Scheduling
Multilevel Queue Scheduling
Multilevel Feedback-Queue Scheduling
 Multi-Processor Scheduling
 OS Examples –Self Study
 Algorithm Evaluation
Algorithm Evaluation
 How do we select the best algorithm?
 Defining a criteria
 CPU Utilization, response time, throughput
 Criteria may include several measures
 e.g. Maximize CPU utilization |and| Response time is 1
 e.g. Maximize Throughput |and| turnaround time
 Analytic Evaluation Method
 The algorithm and some system workload are used to
produce a formula or number which gives the
performance of the algorithm for that workload.




Deterministic modeling
Queuing models
Simulations
Implementation
Deterministic Modeling
 Predetermined workload –defines performance of
each algorithm for that workload. Use of Gantt
charts.
 Simple and fast
 Exact numbers for comparison
 Answers apply only for the cases considered.
 Performance figures may not be true in general
 Suitable: If we are running the same program over
and over again. We can easily select an algorithm.
 Over a set of examples certain trends can be
analyzed, e.g. If all the processes arrive at time 0,
the SJF policy will always result in min waiting time
 If the Arriving time is different the SJF may not always
result in min average waiting time.
Deterministic Modeling
 Process
P1
P2
P3
P4
 Gantt chart
Arrival Time
0.0
2.0
4.0
5.0
P1
0
P2
2
P3
4
Burst Time
7
4
1
4
P2
5
P4
7
P1
11
 Average waiting time = (9 + 1 + 0 +2)/4 = 3
16
Queuing Modeling
 What if the processes that are run, vary from day to day? -Is deterministic method useful?
 The distributions of CPU and I/O bursts can be determined/
approximated/ estimated.
 Mathematical formula: probability of a particular CPU burst
 Arrival times distribution can also be estimated.
 Computer system viewed as a network of queues and
servers: ready queue, I/O queue, event queues, CPUs, I/O
device controllers, etc. e.g. CPU: ready queue, I/O system
:device queue
 Input: Arrival and service rates
 Output: CPU utilization, average queue length, average
waiting time, …
Queuing Modeling
Little’s Formula:
n = λ* W
where
n = average queue length
λ = average arrival rate
W = average waiting time in a
queue
Queuing Modeling
Let the average job arrival rate be 0.5
Algorithm
Average Queue
Length(n)
FCFS
Average Wait
Time
W=tw
4.6
SJF
3.6
1.8
SRTF
3.2
1.6
RR (q=1)
7.0
3.5
RR (q=4)
6.0
3.0
2.3
Queuing Modeling
 Complicated mathematics
 Distributions (uniform, exponential,
etc) for the arrival and departure
rates can be difficult to work with
 Assumptions may not be accurate
 Approximation of the real system
Simulation
 To get more accurate evaluation
 Workload generated by assuming some
distribution and a random number generator,
or by collecting data from the actual system.
 The distributions can be defined
mathematically or empirically.
Simulation
Simulation
 Characteristics
Expensive: hours of
programming and execution
time
May be erroneous because
of the assumptions about
distributions
Implementation
 Even simulation is of limited accuracy
 Best way: Code and implement the scheduling
algorithm and see
 High cost –coding the algorithm
 Modify the OS to support it
 Reaction of the users to changing OS
 Environment changes: if short processes are
given high priority then user may break their
larger processes into many short ones
 For example: Design of an automatic
Interactive/ non-interactive process classifier
Algorithm