Scheduling

Operating Systems: A Modern Perspective, Chapter 7

Slide 7-1

Copyright © 2004 Pearson Education, Inc.


Slide 7-2


7Scheduling


Slide 7-3


CPU Scheduling• Basic Concepts

• Scheduling Criteria

• Scheduling Algorithms– FCFS (FIFO)– SJN– Priority Scheduling– Round Robin Scheduling– Multilevel Queue– Multilevel Feedback Queue


Slide 7-4


CPU Scheduling

• A multiprogramming OS allows more than one process to be loaded in main memory at a time.

• Processes share the CPU using time-multiplexing• A process execution consists of a cycle of CPU

computation--I/O operations.• I/O operations require orders of magnitude more

time to complete.• Basic Idea: When the running process requests an

I/O operation, allocate CPU to another process.


Slide 7-5


CPU Scheduler

• CPU Scheduler: the part of the Process Manager that is responsible for – handling removal of running process from CPU

– Selection of another process

Two major issues:• Scheduling mechanism: how is it all done?• Scheduling policy:

– when is it time for a process to be removed from CPU?

– Which ready process should be allocated the CPU next?


Slide 7-6


Model of Process Execution

ReadyList

ReadyList SchedulerScheduler CPUCPU

ResourceManager

ResourceManager

ResourcesResources

Preemption or voluntary yield

Allocate Request

DoneNewProcess job

jobjob

jobjob

“Ready”“Running”

“Blocked”


Slide 7-7


Scheduler as CPU Resource Manager

Scheduler

ProcessProcess

Units of time for atime-multiplexed CPU

ReleaseReadyto run

Dispatch

Dispatch

Dis

patc

h

Release

Rel

ease

Ready List


Slide 7-8


The Scheduler

Ready Process

EnqueuerEnqueuer ReadyList

ReadyList

DispatcherDispatcher ContextSwitcher

ContextSwitcher

ProcessDescriptor

ProcessDescriptor

CPUCPU

FromOtherStates

Running Process


Slide 7-9


Process/Thread Context

R1R2

Rn

. . .

StatusRegisters

Functional Unit

Left Operand

Right Operand

Result ALU

PC

IRCtl Unit


Slide 7-10


Context Switching

CPU

New Thread Descriptor

Old ThreadDescriptor


Slide 7-11


Scheduling Mechanism CNTD

• When a process is moved to the Ready-List– Process Descriptor (PD) is updated– the enqueuer places a pointer to PD in the Ready-List

• When the Scheduler switches CPU from one process to another process– the Context-Switcher saves the state of the current process in

its PD.

• How context-switching occurs depends on how CPU multiplexing technique used:– voluntary multiplexing– involuntary multiplexing


Slide 7-12


Invoking the Scheduler

• Need a mechanism to call the scheduler

• Voluntary call– Process blocks itself– Calls the scheduler

• Involuntary call– External force (interrupt) blocks the process– Calls the scheduler


Slide 7-13


Voluntary CPU Sharing

yield(pi.pc, pj.pc) { memory[pi.pc] = PC; PC = memory[pj.pc];}

• pi can be “automatically” determined from the processor status registers

yield(*, pj.pc) { memory[pi.pc] = PC; PC = memory[pj.pc];}


Slide 7-14


More on Yield

yield(*, pj.pc);. . .yield(*, pi.pc);. . .yield(*, pj.pc);. . .

• pi and pj can resume one another’s execution

• Suppose pj is the scheduler:// p_i yields to scheduleryield(*, pj.pc);// scheduler chooses pk

yield(*, pk.pc);// pk yields to scheduleryield(*, pj.pc);// scheduler chooses ...


Slide 7-15


Voluntary Sharing

• Every process periodically yields to the scheduler

• Relies on correct process behavior– Malicious– Accidental

• Need a mechanism to override running process


Slide 7-16


Involuntary CPU Sharing• Interval timer

– Device to produce a periodic interrupt– Programmable period

IntervalTimer() { InterruptCount--; if(InterruptCount <= 0) { InterruptRequest = TRUE; InterruptCount = K; }} SetInterval(programmableValue) {

K = programmableValue: InterruptCount = K; }}


Slide 7-17


Contemporary Scheduling

• Involuntary CPU sharing – timer interrupts– Time quantum determined by interval timer –

usually fixed size for every process using the system

– Sometimes called the time slice length


Slide 7-18


Scheduling Mechanism--Dispatcher

• After state of "old" process is saved by context-switcher, the CPU is allocated to the Dispatcher– Dispatcher state is loaded on CPU

• Dispatcher selects one of the ready processes enqueued in the Ready-List.

• Dispatcher performs another context-switch from itself to selected process (saves its state and loads state of selected process).

• The Process Descriptor of selected process is changed from Ready to Running.


Slide 7-19


Choosing a Process to Run

• Mechanism never changes

• Strategy = policy the dispatcher uses to select a process from the ready list

• Different policies for different requirements

Ready Process

EnqueueEnqueue ReadyList

ReadyList

DispatchDispatch ContextSwitch

ContextSwitch

ProcessDescriptor

ProcessDescriptor

CPUCPU

Running Process


Slide 7-20


Policy Considerations

• Policy can control/influence:– CPU utilization– Average time a process waits for service– Average amount of time to complete a job

• Could strive for any of:– Equitability– Favor very short or long jobs– Meet priority requirements– Meet deadlines


Slide 7-21


Optimal Scheduling• Suppose the scheduler knows each process

pi’s service time, pi -- or it can estimate each pi :

• Policy can optimize on any criteria, e.g.,– CPU utilization– Waiting time– Deadline

• To find an optimal schedule:– Have a finite, fixed # of pi

– Know pi for each pi

– Enumerate all schedules, then choose the best


Slide 7-22


However ...

• The (pi) are almost certainly just estimates

• General algorithm to choose optimal schedule is O(n2)

• Other processes may arrive while these processes are being serviced

• Usually, optimal schedule is only a theoretical benchmark – scheduling policies try to approximate an optimal schedule


Slide 7-23


Model of Process Execution

ReadyList


ResourceManager

ResourceManager

ResourcesResources


Allocate Request

DoneNewProcess job

jobjob

jobjob


“Blocked”


Slide 7-24


Simplified Model

ReadyList


ResourceManager

ResourceManager

ResourcesResources

Allocate Request

DoneNewProcess job

jobjob

jobjob


“Blocked”

• Simplified, but still provide analysis result

• Easy to analyze performance

• No issue of voluntary/involuntary sharing



Slide 7-25


Nonpreemptive Schedulers

ReadyList

ReadyList SchedulerScheduler CPUCPU DoneNew

Process

• Try to use the simplified scheduling model

• Only consider running and ready states

• Ignores time in blocked state:– “New process created when it enters ready state”– “Process is destroyed when it enters blocked state”– Really just looking at “small phases” of a process

Blocked or preempted processes


Slide 7-26


Estimating CPU Utilization

ReadyList


Process

System pi per second

Each pi uses 1/ units ofthe CPU

Let = the average rate at which processes are placed in the Ready List, arrival rate

Let = the average service rate 1/ = the average (pi)


Slide 7-27


Estimating CPU Utilization

ReadyList


Process

Let = the average rate at which processes are placed in the Ready List, arrival rate

Let = the average service rate 1/ = the average (pi)

Let = the fraction of the time that the CPU is expected to be busy = # pi that arrive per unit time * avg time each spends on CPU = * 1/ = /

• Notice must have < (i.e., < 1)• What if approaches 1?


Slide 7-28


Talking About Scheduling ...• Let P = {pi | 0 i < n} = set of processes

• Let S(pi) {running, ready, blocked}

• Let (pi) = Time process needs to be in running state (the service time)

• Let W(pi) = Time pi is in ready state before first transition to running (wait time)

• Let TTRnd(pi) = Time from pi first enter ready to last exit ready (turnaround time)

• Batch Throughput rate = inverse of avg TTRnd

• Timesharing response time = W(pi)


Slide 7-29


First-Come-First-Servedi (pi)0 3501 1252 4753 2504 75

p0

TTRnd(p0) = (p0) = 350 W(p0) = 0

0 350


Slide 7-30



p0 p1

TTRnd(p0) = (p0) = 350TTRnd(p1) = ((p1) +TTRnd(p0)) = 125+350 = 475

W(p0) = 0W(p1) = TTRnd(p0) = 350

475350


Slide 7-31



p0 p1 p2

TTRnd(p0) = (p0) = 350TTRnd(p1) = ((p1) +TTRnd(p0)) = 125+350 = 475TTRnd(p2) = ((p2) +TTRnd(p1)) = 475+475 = 950

W(p0) = 0W(p1) = TTRnd(p0) = 350W(p2) = TTRnd(p1) = 475

475 950


Slide 7-32



p0 p1 p2 p3

TTRnd(p0) = (p0) = 350TTRnd(p1) = ((p1) +TTRnd(p0)) = 125+350 = 475TTRnd(p2) = ((p2) +TTRnd(p1)) = 475+475 = 950TTRnd(p3) = ((p3) +TTRnd(p2)) = 250+950 = 1200

W(p0) = 0W(p1) = TTRnd(p0) = 350W(p2) = TTRnd(p1) = 475W(p3) = TTRnd(p2) = 950

1200950


Slide 7-33



p0 p1 p2 p3 p4

TTRnd(p0) = (p0) = 350TTRnd(p1) = ((p1) +TTRnd(p0)) = 125+350 = 475TTRnd(p2) = ((p2) +TTRnd(p1)) = 475+475 = 950TTRnd(p3) = ((p3) +TTRnd(p2)) = 250+950 = 1200TTRnd(p4) = ((p4) +TTRnd(p3)) = 75+1200 = 1275

W(p0) = 0W(p1) = TTRnd(p0) = 350W(p2) = TTRnd(p1) = 475W(p3) = TTRnd(p2) = 950W(p4) = TTRnd(p3) = 1200

1200 1275


Slide 7-34


FCFS Average Wait Time

i (pi)0 3501 1252 4753 2504 75

p0 p1 p2 p3 p4

TTRnd(p0) = (p0) = 350TTRnd(p1) = ((p1) +TTRnd(p0)) = 125+350 = 475TTRnd(p2) = ((p2) +TTRnd(p1)) = 475+475 = 950TTRnd(p3) = ((p3) +TTRnd(p2)) = 250+950 = 1200TTRnd(p4) = ((p4) +TTRnd(p3)) = 75+1200 = 1275

W(p0) = 0W(p1) = TTRnd(p0) = 350W(p2) = TTRnd(p1) = 475W(p3) = TTRnd(p2) = 950W(p4) = TTRnd(p3) = 1200

Wavg = (0+350+475+950+1200)/5 = 2974/5 = 595

127512009004753500

•Easy to implement•Ignores service time, etc•Not a great performer


Slide 7-35


Predicting Wait Time in FCFS

• In FCFS, when a process arrives, all in ready list will be processed before this job

• Let be the service rate

• Let L be the ready list length

• Wavg(p) = L*1/1/L

• Compare predicted wait with actual in earlier examples


Slide 7-36


Shortest Job Nexti (pi)0 3501 1252 4753 2504 75

p4

TTRnd(p4) = (p4) = 75 W(p4) = 0

750


Slide 7-37



p1p4

TTRnd(p1) = (p1)+(p4) = 125+75 = 200

TTRnd(p4) = (p4) = 75

W(p1) = 75

W(p4) = 0

200750


Slide 7-38



p1 p3p4

TTRnd(p1) = (p1)+(p4) = 125+75 = 200

TTRnd(p3) = (p3)+(p1)+(p4) = 250+125+75 = 450TTRnd(p4) = (p4) = 75

W(p1) = 75

W(p3) = 200W(p4) = 0

450200750


Slide 7-39



p0p1 p3p4

TTRnd(p0) = (p0)+(p3)+(p1)+(p4) = 350+250+125+75 = 800TTRnd(p1) = (p1)+(p4) = 125+75 = 200

TTRnd(p3) = (p3)+(p1)+(p4) = 250+125+75 = 450TTRnd(p4) = (p4) = 75

W(p0) = 450W(p1) = 75

W(p3) = 200W(p4) = 0

800450200750


Slide 7-40



p0p1 p2p3p4

TTRnd(p0) = (p0)+(p3)+(p1)+(p4) = 350+250+125+75 = 800TTRnd(p1) = (p1)+(p4) = 125+75 = 200TTRnd(p2) = (p2)+(p0)+(p3)+(p1)+(p4) = 475+350+250+125+75 = 1275TTRnd(p3) = (p3)+(p1)+(p4) = 250+125+75 = 450TTRnd(p4) = (p4) = 75

W(p0) = 450W(p1) = 75W(p2) = 800

W(p3) = 200W(p4) = 0

1275800450200750


Slide 7-41



p0p1 p2p3p4

TTRnd(p0) = (p0)+(p3)+(p1)+(p4) = 350+250+125+75 = 800TTRnd(p1) = (p1)+(p4) = 125+75 = 200TTRnd(p2) = (p2)+(p0)+(p3)+(p1)+(p4) = 475+350+250+125+75 = 1275TTRnd(p3) = (p3)+(p1)+(p4) = 250+125+75 = 450TTRnd(p4) = (p4) = 75

W(p0) = 450W(p1) = 75W(p2) = 800

W(p3) = 200W(p4) = 0

Wavg = (450+75+800+200+0)/5 = 1525/5 = 305

1275800450200750

•Minimizes wait time•May starve large jobs•Must know service times


Slide 7-42


Shortest-Job-Next (SJN) Scheduling

SJN is optimal – gives minimum average waiting time for a given set of processes


Slide 7-43


Preemptive Schedulers

ReadyList



DoneNewProcess

• Highest priority process is guaranteed to be running at all times– Or at least at the beginning of a time slice

• Dominant form of contemporary scheduling

• But complex to build & analyze


Slide 7-44


Example of Non-Preemptive SJN

2.0


Slide 7-45


Example of Preemptive SJN

2.0

Average waiting time = (0 + 0 + 0 + 2)/4 = 0.5


Slide 7-46


Priority Scheduling


Slide 7-47


Priority Schedulingi (pi) Pri0 350 51 125 22 475 33 250 14 75 4

p0p1 p2p3 p4

TTRnd(p0) = (p0)+(p4)+(p2)+(p1) )+(p3) = 350+75+475+125+250 = 1275TTRnd(p1) = (p1)+(p3) = 125+250 = 375TTRnd(p2) = (p2)+(p1)+(p3) = 475+125+250 = 850TTRnd(p3) = (p3) = 250TTRnd(p4) = (p4)+ (p2)+ (p1)+(p3) = 75+475+125+250 = 925

W(p0) = 925W(p1) = 250W(p2) = 375

W(p3) = 0W(p4) = 850

Wavg = (925+250+375+0+850)/5 = 2400/5 = 480

12759258503752500

•Reflects importance of external use•May cause starvation•Can address starvation with aging


Slide 7-48


Deadline Scheduling

• Real Time Systems– Processes must complete their task by specific

deadlines– Main performance criteria– Scheduler must have complete knowledge of

service time of each process• All function must be predictable– no virtual memory

– A process is admitted to ready list only if OS can guarantee deadline can be met.


Slide 7-49


Deadline Schedulingi (pi) Deadline0 350 5751 125 5502 475 10503 250 (none)4 75 200

p0p1 p2 p3p4

12751050550200

0

•Allocates service by deadline•May not be feasible

p0p1 p2 p3p4

p0 p1 p2 p3p4

575


Slide 7-50


Round Robin (RR) Scheduling


Slide 7-51


Round Robin (TQ=50)i (pi)0 3501 1252 4753 2504 75

p0

W(p0) = 0

0 50


Slide 7-52



p0

W(p0) = 0W(p1) = 50

1000p1


Slide 7-53



p0

W(p0) = 0W(p1) = 50W(p2) = 100

1000p2p1


Slide 7-54



p0

W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150

2001000p3p2p1


Slide 7-55



p0

W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150W(p4) = 200

2001000p4p3p2p1


Slide 7-56



p0

W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150W(p4) = 200

3002001000p0p4p3p2p1


Slide 7-57



p0

TTRnd(p4) =

W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150W(p4) = 200

4754003002001000p4p0p4p3p2p1 p1 p2 p3


Slide 7-58



p0

TTRnd(p1) =

TTRnd(p4) =

W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150W(p4) = 200

4754003002001000p4 p1p0p4p3p2p1 p1 p2 p3 p0

550


Slide 7-59



p0

TTRnd(p1) =

TTRnd(p3) = TTRnd(p4) =

W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150W(p4) = 200

4754003002001000p4 p1p0p4p3p2p1 p1 p2 p3 p0 p3p2

p0 p3p2 p0 p3p2

550 650

650 750 850 950


Slide 7-60



p0



W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150W(p4) = 200

4754003002001000p4 p1p0p4p3p2p1 p1 p2 p3 p0 p3p2

p0 p3p2 p0 p3p2 p0 p2 p0

550 650

650 750 850 950 1050


Slide 7-61



p0

TTRnd(p0) = TTRnd(p1) = TTRnd(p2) = TTRnd(p3) = TTRnd(p4) =

W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150W(p4) = 200

4754003002001000p4 p1p0p4p3p2p1 p1 p2 p3 p0 p3p2

p0 p3p2 p0 p3p2 p0 p2 p0 p2 p2 p2 p2

550 650

650 750 850 950 1050 1150 1250 1275


Slide 7-62



p0


W(p0) = 0W(p1) = 50W(p2) = 100W(p3) = 150W(p4) = 200

Wavg = (0+50+100+150+200)/5 = 500/5 = 100

4754003002001000

•Equitable•Most widely-used•Fits naturally with interval timer

p4 p1p0p4p3p2p1 p1 p2 p3 p0 p3p2

p0 p3p2 p0 p3p2 p0 p2 p0 p2 p2 p2 p2

550 650

650 750 850 950 1050 1150 1250 1275

TTRnd_avg = (1100+550+1275+950+475)/5 = 4350/5 = 870


Slide 7-63


RR with Overhead=10 (TQ=50)i (pi)0 3501 1252 4753 2504 75

p0


W(p0) = 0W(p1) = 60W(p2) = 120W(p3) = 180W(p4) = 240

Wavg = (0+60+120+180+240)/5 = 600/5 = 120

5404803602401200

•Overhead must be considered

p4 p1p0p4p3p2p1 p1 p2 p3 p0 p3p2

p0 p3p2 p0 p3p2 p0 p2 p0 p2 p2 p2 p2

575 790

910 1030 1150 1270 1390 1510 1535

TTRnd_avg = (1320+660+1535+1140+565)/5 = 5220/5 = 1044

635 670

790


Slide 7-64


Multi-Level Queue• Extension of priority scheduling, which also

combines other strategies• Ready list is partitioned into multiple sub-lists• Each process is assigned to a queue based on some

criteria (type, priority, etc.)– E.g. one Q for foreground processes and one for

background processes

• Scheduler: * in-queue strategy * cross-queue strategy

– E.g. In-queue strategy: RR for foreground Q and FCFS for background A

– Cross-queue strategy: Serve higher-level processes first


Slide 7-65


Multi-Level Queue

• Example: – Cross-Queue Strategy: Each queue get a percentage of

CPU time (in a round robin fashion) which it can schedule among its processes.

– E.g. Level 1: 80% of CPU time

Level 2: 20% of CPU time


Slide 7-66


Multi-Level Queues

Ready List0Ready List0




SchedulerScheduler CPUCPU


DoneNewProcess

•All processes at level i run before any process at level j•At a level, use another policy, e.g. RR


Slide 7-67


Multi-Level Feedback Queue


Slide 7-68


Example:Multi-Level Feedback Queue

• Gives shorter jobs higher priority without needing to predict a job’s service time requirement.


Slide 7-69


Contemporary Scheduling• Involuntary CPU sharing -- timer interrupts

– Time quantum determined by interval timer -- usually fixed for every process using the system

– Sometimes called the time slice length

• Priority-based process (job) selection– Select the highest priority process– Priority reflects policy

• With preemption

• Usually a variant of Multi-Level Queues


Slide 7-70


BSD 4.4 Scheduling

• Involuntary CPU Sharing

• Preemptive algorithms

• 32 Multi-Level Queues– Queues 0-7 are reserved for system functions– Queues 8-31 are for user space functions– nice influences (but does not dictate) queue

level


Slide 7-71


Windows NT/2K Scheduling

• Involuntary CPU Sharing across threads

• Preemptive algorithms

• 32 Multi-Level Queues– Highest 16 levels are “real-time”– Next lower 15 are for system/user threads

• Range determined by process base priority

– Lowest level is for the idle thread


Slide 7-72


Bank Teller Simulation

Tellers at the Bank

T1T1

T2T2

TnTn

…

Model of Tellers at the Bank

Cus

tom

ers

Arr

ival

s


Slide 7-73


Simulation Kernel Loop

simulated_time = 0;while (true) {

event = select_next_event();if (event->time > simulated_time)

simulated_time = event->time;evaluate(event->function, …);

}


Slide 7-74


Simulation Kernel Loop(2)void runKernel(int quitTime){ Event *thisEvent; // Stop by running to elapsed time, or by causing quit execute if(quitTime <= 0) quitTime = 9999999; simTime = 0; while(simTime < quitTime) { // Get the next event if(eventList == NIL) { // No more events to process break; } thisEvent = eventList; eventList = thisEvent->next; simTime = thisEvent->getTime(); // Set the time // Execute this event thisEvent->fire(); delete(thisEvent); };}

Scheduling

Documents

Transcript of Scheduling