0% found this document useful (0 votes)
68 views

CS204-OS-module 4

The document discusses various CPU scheduling algorithms. It begins by explaining the CPU-I/O burst cycle where processes alternate between CPU execution and I/O wait states. It then discusses the role of the CPU scheduler in selecting processes from ready queues. Common scheduling criteria like CPU utilization, throughput, turnaround time and waiting time are outlined. Specific algorithms covered include first-come, first-served (FCFS), shortest job first (SJF), and priority scheduling. SJF is shown to provide optimal scheduling with minimum average waiting times.

Uploaded by

Abitha Sajan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
68 views

CS204-OS-module 4

The document discusses various CPU scheduling algorithms. It begins by explaining the CPU-I/O burst cycle where processes alternate between CPU execution and I/O wait states. It then discusses the role of the CPU scheduler in selecting processes from ready queues. Common scheduling criteria like CPU utilization, throughput, turnaround time and waiting time are outlined. Specific algorithms covered include first-come, first-served (FCFS), shortest job first (SJF), and priority scheduling. SJF is shown to provide optimal scheduling with minimum average waiting times.

Uploaded by

Abitha Sajan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

1

MODULE IV

CPU SCHEDULING
Basic concepts
 Maximum CPU utilization obtained with multiprogramming

CPU-I/O Burst Cycle


 Process execution consists of a cycle of CPU execution and I/0 wait. Processes
alternate between these two states.
 Process execution begins with a CPU burst.
 Final CPU burst ends with a system request to terminate execution.

KTU STUDENTS

Alternating sequence of CPU and 1/0 bursts.


 An I/O-bound program typically has many short CPU bursts. A CPU-bound
program might have a few long CPU bursts.
 This distribution can be important in the selection of an appropriate CPU-
scheduling algorithm
CPU Scheduler
 Whenever the CPU becomes idle, the operating system must select one of the
processes in the ready queue to be executed. The selection process is carried out
by the short-term scheduler (or CPU scheduler).
 The scheduler selects a process from the processes in memory that are ready to
execute and allocates the CPU to that process.

For more study materials>www.ktustudents.in


2

 A ready queue can be implemented as a FIFO queue, a priority queue, a tree, or


sirnply an unordered linked list.
 The records in the queues are generally process control blocks (PCBs) of the
processes.

Preemptive Scheduling

 CPU-scheduling decisions may take place under the following four


circumstances:
1. When a process switches from the running state to the waiting state
2. When a process switches from the running state to the ready state.
3. When a process switches from the waiting state to the ready state
4. When a process terminates
 Scheduling under 1 and 4 is nonpreemptive.
 All other scheduling is preemptive
 Consider access to shared data
 Consider preemption while in kernel mode
 Consider interrupts occurring during crucial OS activities
Dispatcher
 Dispatcher module gives control of the CPU to the process selected by the short-
KTU STUDENTS
term scheduler; this involves:
1. Switching context
2. Switching to user mode
3. Jumping to the proper location in the user program to restart that
program
 The time it takes for the dispatcher to stop one process and start another running
is known as the dispatch latency.

Scheduling Criteria
 Many criteria have been suggested for comparing CPU-scheduling algorithms.
 The criteria include the following:
1. CPU utilization- Keep the CPU as busy as possible.
2. Throughput – Number of processes that complete their execution per
time unit
3. Turnaround Time-The interval from the time of submission of a process
to the time of completion. Turnaround Time is the sum of the periods
spent waiting to get into memory, waiting in the ready queue, executing
on the CPU, and doing I/0 waiting time.
4. Waiting Time-sum of the periods spent waiting in the ready queue.

For more study materials>www.ktustudents.in


3

5. 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)

SCHEDULING ALGORITHMS

First- Come, First-Served (FCFS) Scheduling


 The process that requests the CPU first is allocated the CPU first.
 The implementation of the FCFS policy is easily managed with a FIFO queue.
 When a process enters the ready queue, its PCB is linked onto the tail of the queue.
 When the CPU is free, it is allocated to the process at the head of the queue.
 The running process is then removed from the queue.

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:

0
KTU STUDENTS
P1

 Waiting time for P1 = 0; P2 = 24; P3 = 27


24
P2

27
P3

30

 Average waiting time: (0 + 24 + 27)/3 = 17


 Suppose that the processes arrive in the order: P2 , P3 , P1
 The Gantt chart for the schedule is:
P2 P3 P1

0 3 6 30

 Waiting time for P1 = 6; P2 = 0; P3 = 3


 Average waiting time: (6 + 0 + 3)/3 = 3
 Convoy effect -All the processes wait for the one big process to get off the CPU.
 FCFS scheduling algorithm is Non-preemptive.

Shortest-Job-First Scheduling

 This algorithm associates with each process the length of the process's next CPU
burst. When the CPU is available, it is assigned to the process that has the
smallest next CPU burst.
 More appropriate term for this scheduling method would be the shortest-next-
CPU-burst algorithm.

For more study materials>www.ktustudents.in


4

Process Burst Time


P1 6
P2 8
P3 7
P4 3
 SJF scheduling chart:

P4 P1 P3 P2

0 3 9 16 24

 Average waiting time = (3 + 16 + 9 + 0) / 4 = 7


 SJF is optimal – Gives minimum average waiting time for a given set of
processes.
 The difficulty is knowing the length of the next CPU request.
 Can only estimate the length – should be similar to the previous one
 Then pick process with shortest predicted next CPU burst
 Can be done by using the length of previous CPU bursts, using exponential
averaging

1. t n  actual
th
length of n CPU burst

KTU STUDENTS
2.  n  1  predicted

3.  , 0    1
value for the next CPU burst

4. Define :

 Commonly, α set to ½
 Preemptive version called shortest-remaining-time-first
  =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
 A preemptive SJF algorithm will preempt the currently executing process,
whereas a nonpreemptive SJF algorithm will allow the currently running process
to finish its CPU burst.

For more study materials>www.ktustudents.in


5

 Preemptive SJF scheduling is sometimes called shortest-remaining-time-first


scheduling.
 Now we add the concepts of varying arrival times and preemption to the analysis
Process Arrival Time Burst Time

P1 0 8
P2 1 4
P3 2 9
P4 3 5

 Preemptive SJF Gantt Chart

P1 P2 P4 P1 P3

0 1 5 10 17 26

 Average waiting time = [(10-1)+(1-1)+(17-2)+5-3)]/4 = 26/4 = 6.5 msec

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
KTU STUDENTS
 Nonpreemptive

ProcessA Burst Time Priority

P1 10 3
P2 1 1
P3 2 4
P4 1 5
P5 5 2

 Priority scheduling Gantt Chart


P1 P2 P1 P3 P4

0 1 6 16 18 19

 Average waiting time = 8.2 msec


 Priorities can be defined either internally or externally.

For more study materials>www.ktustudents.in


6

 A rnajor problem with priority scheduling algorithms is indefinite blocking, or


starvation, leaving low priority processes waiting indefinitely.
 Solution for starvation is Aging, a technique of gradually increasing the priority
of processes that wait in the system for a long time.

Round-Robin Scheduling
 The round-robin (RR) scheduling algorithm is designed especially for
timesharing systems.
 Each process gets a small unit of CPU time (time quantum q), 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.
 Timer interrupts every quantum to schedule next process
 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 = 4
Process Burst Time
P1
P2
KTU STUDENTS
24
3
P3 3
 The Gantt chart is:

P1 P2 P3 P1 P1 P1 P1 P1

0 4 7 10 14 18 22 26 30

 Typically, higher average turnaround than SJF, but better response


 q should be large compared to context switch time
 q usually 10ms to 100ms, context switch < 10 microsecond

 Turnaround time also depends on the size of the time quantum.

For more study materials>www.ktustudents.in


7

 In general, the average turnaround time can be improved if most processes finish
their next CPU burst in a single time quantum.

Multilevel Queue Scheduling

 In this scheduling processes are classified into different groups


 Ready queue is partitioned into separate queues,
 foreground (interactive)
 background (batch)
 These two types of processes have different response-time requirements and so
may have different scheduling needs. In addition, foreground processes may
have priority (externally defined) over background processes.
 A multilevel queue scheduling algorithm partitions the ready queue into several
separate queues. The processes are permanently assigned to one queue
 Each queue has its own scheduling algorithm:
 foreground – RR
 background – FCFS

KTU STUDENTS

 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 Feedback Queue


 The multilevel feedback queue scheduling algorithm, allows a process to move
between 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

For more study materials>www.ktustudents.in


8

 method used to determine which queue a process will enter when that
process needs service

 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

Deadlocks
KTU STUDENTS
A process requests resources; if the resources are not available at that time, the process
enters a waiting state. Sometimes, a waiting process is never again able to change state,
because the resources it has requested are held by other waiting processes. This
situation is called a deadlock.

System Model

A system consists of a finite number of resources to be distributed among a number of


competing processes. The resources may be partitioned into several types (or classes),
each consisting of some number of identical instances. CPU cycles, files, and I/O
devices (such as printers and DVD drives) are examples of resource types. If a system
has two CPUs, then the resource type CPU has two instances. Similarly, the resource
type printer may have five instances.

Under the normal mode of operation, a process may utilize a resource in only the
following sequence:
1. Request: The process requests the resource. If the request cannot be granted
immediately (for example, if the resource is being used by another process), then the
requesting process must wait until it can acquire the resource.
2. Use: The process can operate on the resource (for example, if the resource
is a printer, the process can print on the printer).

For more study materials>www.ktustudents.in


9

3. Release: The process releases the resource.

Deadlock Characterization

Deadlock can arise if four conditions hold simultaneously


 Mutual exclusion: only one process at a time can use a resource

 Hold and wait: a process holding at least one resource is waiting to acquire
additional resources held by other processes

 No preemption: a resource can be released only voluntarily by the process


holding it, after that process has completed its task

 Circular wait: there exists a set {P0, P1, …, Pn} of waiting processes such that P0
is waiting for a resource that is held by P1, P1 is waiting for a resource that is
held by P2, …, Pn–1 is waiting for a resource that is held by Pn, and Pn is waiting
for a resource that is held by P0.

Resource-Allocation Graph

Deadlocks can be described more precisely in terms of a directed graph called a system
resource-allocation graph. A set of vertices V and a set of edges E.


KTU STUDENTS
 V is partitioned into two types:

P = {P1, P2, …, Pn}, the set consisting of all the processes in the system

 R = {R1, R2, …, Rm}, the set consisting of all resource types in the system

 request edge – directed edge Pi  Rj

 assignment edge – directed edge Rj  Pi

 Process

 Resource Type with 4 instances

 Pi requests instance of Rj

 Pi is holding an instance of Rj

For more study materials>www.ktustudents.in


10

Example of a Resource Allocation Graph

Resource Allocation Graph With A Deadlock

KTU STUDENTS

Since no resource instance is currently available, we add a request edge P3→ R2 to the
above graph .At this point, two minimal cycles exist in the system:
P1 → R1 → P2 → R3 → P3 → R2 → P1
P2 → R3 → P3 → R2 → P2
Processes P1, P2, and P3 are deadlocked. Process P2 is waiting for the resource R3,
which is held by process P3. Process P3 is waiting for either process P1 or process P2 to
release resource R2. In addition, process P1 is waiting for process P2 to release resource
R1.

Graph With A Cycle But No Deadlock

Now consider the resource-allocation graph below. In this example, we also have a
cycle:
P1 → R1 → P3 → R2 → P1

For more study materials>www.ktustudents.in


11

However, there is no deadlock. Observe that process P4 may release its instance of
resource type R2. That resource can then be allocated to P3, breaking the cycle.

Basic Facts

 If graph contains no cycles  no deadlock

 If graph contains a cycle 

 if only one instance per resource type, then deadlock

 KTU STUDENTS
if several instances per resource type, possibility of deadlock

Methods for Handling Deadlocks

 Ensure that the system will never enter a deadlock state:

 Deadlock prevention

 Deadlock avoidance

 Allow the system to enter a deadlock state and then recover

 Ignore the problem and pretend that deadlocks never occur in the system; used
by most operating systems, including UNIX

Deadlock Prevention

Restrain the ways request can be made

 Mutual Exclusion – not required for sharable resources (e.g., read-only files);
must hold for non-sharable resources

For more study materials>www.ktustudents.in


12

 Hold and Wait – must guarantee that whenever a process requests a resource, it
does not hold any other resources

 Require process to request and be allocated all its resources before it


begins execution, or allow process to request resources only when the
process has none allocated to it.

 Low resource utilization; starvation possible

 No Preemption –

 If a process that is holding some resources requests another resource that


cannot be immediately allocated to it, then all resources currently being
held are released

 Preempted resources are added to the list of resources for which the
process is waiting

 Process will be restarted only when it can regain its old resources, as well
as the new ones that it is requesting

 Circular Wait – impose a total ordering of all resource types, and require that
each process requests resources in an increasing order of enumeration

Deadlock Avoidance KTU STUDENTS


Requires that the system has some additional a priori information available

 Simplest and most useful model requires that each process declare the maximum
number of resources of each type that it may need

 The deadlock-avoidance algorithm dynamically examines the resource-allocation


state to ensure that there can never be a circular-wait condition

 Resource-allocation state is defined by the number of available and allocated


resources, and the maximum demands of the processes

Safe State

 When a process requests an available resource, system must decide if immediate


allocation leaves the system in a safe state

 System is in safe state if there exists a sequence <P1, P2, …, Pn> of ALL the
processes in the systems such that for each Pi, the resources that Pi can still
request can be satisfied by currently available resources + resources held by all
the Pj, with j < I

For more study materials>www.ktustudents.in


13

 That is:

 If Pi resource needs are not immediately available, then Pi can wait until
all Pj have finished

 When Pj is finished, Pi can obtain needed resources, execute, return


allocated resources, and terminate

 When Pi terminates, Pi +1 can obtain its needed resources, and so on

 If a system is in safe state  no deadlocks

 If a system is in unsafe state  possibility of deadlock

 Avoidance  ensure that a system will never enter an unsafe state.

Safe, Unsafe, Deadlock State

KTU STUDENTS

Avoidance Algorithms

 Single instance of a resource type

 Use a resource-allocation graph

 Multiple instances of a resource type

 Use the banker’s algorithm

For more study materials>www.ktustudents.in


14

Resource-Allocation Graph Scheme

 Claim edge Pi  Rj indicated that process Pj may request resource Rj; represented
by a dashed line

 Claim edge converts to request edge when a process requests a resource

 Request edge converted to an assignment edge when the resource is allocated to


the process

 When a resource is released by a process, assignment edge reconverts to a claim


edge

 Resources must be claimed a priori in the system

KTU STUDENTS

Unsafe State In Resource-Allocation Graph

For more study materials>www.ktustudents.in


15

Resource-Allocation Graph Algorithm

 Suppose that process Pi requests a resource Rj

 The request can be granted only if converting the request edge to an assignment
edge does not result in the formation of a cycle in the resource allocation graph

Banker’s Algorithm
 Multiple instances

 Each process must a priori claim maximum use

 When a process requests a resource it may have to wait

 When a process gets all its resources it must return them in a finite amount of
time

Data Structures for the Banker’s Algorithm

Let n = number of processes, and m = number of resources types.

 Available: Vector of length m. If available [j] = k, there are k instances of


resource type Rj available

KTU STUDENTS
 Max: n x m matrix. If Max [i,j] = k, then process Pi may request at most k
instances of resource type Rj

 Allocation: n x m matrix. If Allocation[i,j] = k then Pi is currently allocated k


instances of Rj

 Need: n x m matrix. If Need[i,j] = k, then Pi may need k more instances of Rj to


complete its task

Need [i,j] = Max[i,j] – Allocation [i,j]

Safety Algorithm

1 Let Work and Finish be vectors of length m and n, respectively. Initialize:

Work = Available

Finish [i] = false for i = 0, 1, …, n- 1

2. Find an i such that both:

(a) Finish [i] = false

(b) Needi  Work

For more study materials>www.ktustudents.in


16

If no such i exists, go to step 4

3. Work = Work + Allocationi


Finish[i] = true
go to step 2

4. If Finish [i] == true for all i, then the system is in a safe state

Resource-Request Algorithm for Process Pi

Requesti = request vector for process Pi. If Requesti [j] = k then process Pi wants k
instances of resource type Rj

1. If Requesti  Needi go to step 2. Otherwise, raise error condition, since


process has exceeded its maximum claim

2. If Requesti  Available, go to step 3. Otherwise Pi must wait, since


resources are not available

3. Pretend to allocate requested resources to Pi by modifying the state as


follows:

Available = Available – Requesti;


KTU STUDENTS
Allocationi = Allocationi + Requesti;

Needi = Needi – Requesti;

o If safe  the resources are allocated to Pi


o If unsafe  Pi must wait, and the old resource-allocation state is
restored

Example of Banker’s Algorithm

 5 processes P0 through P4;

 3 resource types:

A (10 instances), B (5instances), and C (7 instances)

 Snapshot at time T0:

For more study materials>www.ktustudents.in


17

Allocation Max Available

ABC ABC ABC

P0 010 753 332

P1 200 322

P2 302 902

P3 211 222

P4 002 433

 The content of the matrix Need is defined to be Max – Allocation

Need

ABC

P0 743

P1 122

P2 600

P3 KTU STUDENTS
011

P4 431

 The system is in a safe state since the sequence < P1, P3, P4, P2, P0> satisfies safety
criteria

Example: P1 Request (1,0,2)

 Check that Request  Available (that is, (1,0,2)  (3,3,2)  true

Allocation Need Available

ABC ABC ABC

P0 010 743 230

P1 302 020

P2 302 600

P3 211 011

P4 002 431

For more study materials>www.ktustudents.in


18

 Executing safety algorithm shows that sequence < P1, P3, P4, P0, P2> satisfies
safety requirement

Deadlock Detection

 Allow system to enter deadlock state

 Detection algorithm

 Recovery scheme

Single Instance of Each Resource Type

 Maintain wait-for graph

l Nodes are processes

l Pi  Pj if Pi is waiting for Pj

 Periodically invoke an algorithm that searches for a cycle in the graph. If there is
a cycle, there exists a deadlock

 An algorithm to detect a cycle in a graph requires an order of n2 operations,


where n is the number of vertices in the graph

KTU STUDENTS
Resource-Allocation Graph and Wait-for Graph

Resource-Allocation Graph Corresponding wait-for graph

Several Instances of a Resource Type

For more study materials>www.ktustudents.in


19

 Available: A vector of length m indicates the number of available resources of


each type

 Allocation: An n x m matrix defines the number of resources of each type


currently allocated to each process

 Request: An n x m matrix indicates the current request of each process. If


Request [i][j] = k, then process Pi is requesting k more instances of resource type
Rj.

Detection Algorithm

1 Let Work and Finish be vectors of length m and n, respectively Initialize:

(a) Work = Available

(b) For i = 1,2, …, n, if Allocationi  0, then


Finish[i] = false; otherwise, Finish[i] = true

2. Find an index i such that both:

(a) Finish[i] == false

(b) Requesti  Work

KTU STUDENTS
If no such i exists, go to step 4

3. Work = Work + Allocationi


Finish[i] = true
go to step 2

4. If Finish[i] == false, for some i, 1  i  n, then the system is in deadlock state.


Moreover, if Finish[i] == false, then Pi is deadlocked

Example of Detection Algorithm

 Five processes P0 through P4; three resource types


A (7 instances), B (2 instances), and C (6 instances)

 Snapshot at time T0:

Allocation Request Available

ABC ABC ABC

P0 010 000 000

P1 200 202

For more study materials>www.ktustudents.in


20

P2 303 000

P3 211 100

P4 002 002

 Sequence <P0, P2, P3, P1, P4> will result in Finish[i] = true for all i

 P2 requests an additional instance of type C

Request

ABC

P0 000

P1 202

P2 001

P3 100

P4 002

 State of system

 KTU STUDENTS
Can reclaim resources held by process P0, but insufficient resources to
fulfill other processes; requests

 Deadlock exists, consisting of processes P1, P2, P3, and P4\

Recovery from Deadlock: Process Termination

 Abort all deadlocked processes

 Break the deadlock cycle

 But at grant expense

 Abort one process at a time until the deadlock cycle is eliminated

 After each process is aborted deadlock detection algorithm must be invoked to


determine whether any process still in dead lock.

 The order should we choose to abort

1. Priority of the process

2. How long process has computed, and how much longer to completion?

For more study materials>www.ktustudents.in


21

3. Resources the process has used

4. Resources process needs to complete

5. How many processes will need to be terminated

6. Is process interactive or batch

Recovery from Deadlock: Resource Preemption

 Selecting a victim – minimize cost

 Which process and which resources are to be preempted.

 We must determine the order of preemption to minimize cost.

 Rollback – return to some safe state, restart process for that state

 Starvation – same process may always be picked as victim, include number of


rollback in cost factor

KTU STUDENTS

Prapared By: Shahad P., Jasheeda P

Assistant Professors,

Department of CSE,

MEA ENGINEERING COLLEGE,

PERINTHALMANNA

For more study materials>www.ktustudents.in

You might also like