Lesson 4:

CPU Scheduling
 Outline
2

 Basic Concepts
 Scheduling Criteria
 Scheduling Algorithms
 Thread Scheduling
 Multi-Processor Scheduling
 Real-Time CPU Scheduling
 Objectives
3

 Describe various CPU scheduling algorithms

 Assess CPU scheduling algorithms based on scheduling criteria
 Explain the issues related to multiprocessor and multicore
scheduling
 Describe Real-time CPU scheduling
 Basic Concepts
4

 Maximum CPU utilization obtained

with multiprogramming
 CPU–I/O Burst Cycle – Process
execution consists of a cycle of
CPU execution and I/O wait
 CPU burst followed by I/O burst
 CPU burst distribution is of main
concern
 Histogram of CPU-burst Times
5
Large number of short bursts

Small number of longer bursts

 CPU Scheduler
6
 The CPU scheduler selects from among the processes in ready queue, and allocates
a CPU core to one of them
 Queue may be ordered in various ways

 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

 For situations 1 and 4, there is no choice in terms of scheduling. A new process (if one
exists in the ready queue) must be selected for execution.
 For situations 2 and 3, however, there is a choice.
 Preemptive and Nonpreemptive Scheduling
7

 When scheduling takes place only under circumstances 1 and

4, the scheduling scheme is nonpreemptive.
 Otherwise, it is preemptive.
 Under Nonpreemptive scheduling, once the CPU has been
allocated to a process, the process keeps the CPU until it
releases it either by terminating or by switching to the waiting
state.
 Virtually all modern operating systems including Windows,
MacOS, Linux, and UNIX use preemptive scheduling algorithms.
 Preemptive Scheduling and Race Conditions
8

 Preemptive scheduling can result in race conditions when

data are shared among several processes.
 Consider the case of two processes that share data. While
one process 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.
 This issue will be explored in detail in Lesson 5.
 Dispatcher
9

 Dispatcher module gives control of the

CPU to the process selected by the
CPU 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
 Scheduling Criteria
10

 CPU utilization – keep the CPU as busy as possible

 Throughput – # of processes that complete their execution per
time unit
 Turnaround time – amount of time to execute a particular process
 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.
 Scheduling Algorithm Optimization Criteria
11

 Max CPU utilization

 Max throughput
 Min turnaround time
 Min waiting time
 Min response time
 First- Come, First-Served (FCFS) Scheduling
12

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 P2 P3
0 24 27 30

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

 Average waiting time: (0 + 24 + 27)/3 = 17
 FCFS Scheduling (Cont.)
13
Suppose that the processes arrive in the order:
P2 , P 3 , 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
 Much better than previous case
 Convoy effect - short process behind long process
 Consider one CPU-bound and many I/O-bound processes
 Shortest-Job-First (SJF) Scheduling
14

 Associate with each process the length of its next CPU

burst
 Use these lengths to schedule the process with the shortest
time
 SJF is optimal – gives minimum average waiting time for a
given set of processes
 Preemptive version called shortest-remaining-time-first
 How do we determine the length of the next CPU burst?
 Could ask the user
 Estimate
 Example of SJF
15
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

 Determining Length of Next CPU Burst
16

 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

 Commonly, α set to ½
 Prediction of the Length of the Next CPU Burst
17
 Examples of Exponential Averaging
18

  =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

successor predecessor term has less weight than its
predecessor
 Shortest Remaining Time First (SRTF)Scheduling
19

 SRTF is the preemptive version of SJF. The next CPU burst of

the newly arrived process may be shorter than what is left of
the currently executing process, the SRTF algorithm will
preempt the currently executing process.
 Whenever a new process arrives in the ready queue, the
decision on which process to schedule next is redone using
the SRTF algorithm.
 Is SRTF more “optimal” than SJF in terms of the minimum
average waiting time for a given set of processes?
 Example of Shortest-remaining-time-first
20
 Now we add the concepts of varying arrival times and
preemption to the analysis
Process i Arrival TimeT 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

 Round Robin (RR)
21

 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 (FCFS)
 q small  RR
 Note that q must be large with respect to context switch,
otherwise overhead is too high
 Example of RR with Time Quantum = 4
22

Process Burst Time

P1 24
P2 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 10 milliseconds to 100 milliseconds,
 Context switch < 10 microseconds
 Time Quantum and Context Switch Time
23
 Priority Scheduling
24

 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
 Nonpreemptive

 SJF is priority scheduling where priority is the inverse of predicted

next CPU burst time

 Problem  Starvation – low priority processes may never execute

 Solution  Aging – as time progresses increase the priority of the

process
 Example of Priority Scheduling
25
Process Burst Time Priority
P1 10 3
P2 1 1
P3 2 4
P4 1 5
P5 5 2

 Priority scheduling Gantt Chart

 Average waiting time = 8.2

 Priority Scheduling w/ Round-Robin
26
 Run the process with the highest priority. Processes with
the same priority run round-robin
 Example:
Process Burst Time Priority
P1 4 3
P2 5 2
P3 8 2
P4 7 1
P5 3 3
 Gantt Chart with time quantum = 2
 Multilevel Queue
27

 The ready queue consists of multiple queues

 Multilevel queue scheduler defined by the following
parameters:
 Number of queues
 Scheduling algorithms for each queue
 Method used to determine which queue a process will enter
when that process needs service
 Scheduling among the queues
 Multilevel Queue(Cont.)
28
 With priority scheduling, have separate queues for each
priority.
 Schedule the process in the highest-priority queue!
 Multilevel Queue(Cont.)
29

 Prioritization based upon process type

 Multilevel Feedback Queue
30

 A process can move between the various queues.

 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
 Aging can be implemented using multilevel feedback queue
 Example of Multilevel Feedback Queue
31

 Three queues:
 Q0 – RR with time quantum 8
milliseconds
 Q1 – RR time quantum 16 milliseconds
 Q2 – FCFS
 Scheduling
 A new process enters queue Q0 which
is served in RR
 When it gains CPU, the process
receives 8 milliseconds
 If it does not finish in 8 milliseconds, the
process is moved to queue Q1
 At Q1 job is again served in RR and
receives 16 additional milliseconds
 If it still does not complete, it is
preempted and moved to queue Q2
 Thread Scheduling
32

 Distinction between user-level and kernel-level threads

 When threads supported, threads scheduled, not processes
 Many-to-one and many-to-many models, thread library
schedules user-level threads to run on LWP
 Known as process-contention scope (PCS) since scheduling
competition is within the process
 Typically done via priority set by programmer
 Kernel thread scheduled onto available CPU is system-
contention scope (SCS) – competition among all threads in
system
 Pthread Scheduling
33

 API allows specifying either PCS or SCS during thread creation

 PTHREAD_SCOPE_PROCESS schedules threads using PCS scheduling
 PTHREAD_SCOPE_SYSTEM schedules threads using SCS scheduling
 Can be limited by OS – Linux and macOS only allow
PTHREAD_SCOPE_SYSTEM
 Pthread Scheduling API
34
#include <pthread.h>
#include <stdio.h>
#define NUM_THREADS 5
int main(int argc, char *argv[]) {
int i, scope;
pthread_t tid[NUM THREADS];
pthread_attr_t attr;
/* get the default attributes */
pthread_attr_init(&attr);
/* first inquire on the current scope */
if (pthread_attr_getscope(&attr, &scope) != 0)
fprintf(stderr, "Unable to get scheduling scope\n");
else {
if (scope == PTHREAD_SCOPE_PROCESS)
printf("PTHREAD_SCOPE_PROCESS");
else if (scope == PTHREAD_SCOPE_SYSTEM)
printf("PTHREAD_SCOPE_SYSTEM");
else
fprintf(stderr, "Illegal scope value.\n");
}
 Pthread Scheduling API
35

/* set the scheduling algorithm to PCS or SCS */

pthread_attr_setscope(&attr, PTHREAD_SCOPE_SYSTEM);
/* create the threads */
for (i = 0; i < NUM_THREADS; i++)
pthread_create(&tid[i],&attr,runner,NULL);
/* now join on each thread */
for (i = 0; i < NUM_THREADS; i++)
pthread_join(tid[i], NULL);
}
/* Each thread will begin control in this function */
void *runner(void *param)
{
/* do some work ... */
pthread_exit(0);
}
 Multiple-Processor Scheduling
36

 CPU scheduling more complex when multiple CPUs are available

 Multiprocess may be any one of the following architectures:
 Multicore CPUs
 Multithreaded cores
 NUMA (non-uniform memory access) systems
 Heterogeneous multiprocessing
 Multiple-Processor Scheduling
37

 Symmetric multiprocessing (SMP) is where each

processor is self scheduling.
 All threads may be in a common ready queue (a)
 Each processor may have its own private queue of
threads (b)
 Multicore Processors
38

 Recent trend to place multiple processor cores on same

physical chip
 Faster and consumes less power
 Multiple threads per core also growing
 Takes advantage of memory stall to make progress on another
thread while memory retrieve happens
 Multithreaded Multicore System
39

 Each core has > 1 hardware threads.

 If one thread has a memory stall, switch to another thread!
 Multithreaded Multicore System(Cont.)
40

 Chip-multithreading (CMT)
assigns each core multiple
hardware threads. (Intel refers
to this as hyperthreading.)

 On a quad-core system with 2

hardware threads per core,
the operating system sees 8
logical processors.
 Multithreaded Multicore System (Cont.)
41

 Two levels of scheduling:

1. The operating system

deciding which
software thread to run
on a logical CPU

2. How each core decides

which hardware thread
to run on the physical
core.
 Real-Time CPU Scheduling
42

 Can present obvious challenges

 Soft real-time systems – Critical real-time tasks have the highest
priority, but no guarantee as to when tasks will be scheduled
 Hard real-time systems – task must be serviced by its deadline
 Real-Time CPU Scheduling: Latency
43

 Event latency – the amount

of time that elapses from
when an event occurs to
when it is serviced.
 Two types of latencies affect
performance
1. Interrupt latency – time
from arrival of interrupt to
start of routine that services
interrupt
2. Dispatch latency – time for
schedule to take current
process off CPU and switch
to another
 Interrupt Latency
44
 Dispatch Latency
45

 Conflict phase of
dispatch latency:
1. Preemption of
any process
running in kernel
mode
2. Release by low-
priority process
of resources
needed by high-
priority processes
 Homework 1
46

 Read and report:

 Linux scheduling
 Windows scheduling
 Homework 2
47

 Deploying program in slides 34, 35

 Homework 3
48

 Do exercises 5.17 and 5.18 in the textbook

End of Lesson 4