OS Spring’04

Processes

Operating Systems Spring 2004

OS Spring’04

What is a process? An instance of an application

execution Process is the most basic

abstractions provided by OSAn isolated computation context for each application

Computation contextCPU state + address space + environment

OS Spring’04

CPU state=Register contents Process Status Word (PSW)

exec. mode, last op. outcome, interrupt level

Instruction Register (IR)Current instruction being executed

Program counter (PC) Stack pointer (SP) General purpose registers

OS Spring’04

Address space Text

Program code

DataPredefined data (known in compile time)

HeapDynamically allocated data

StackSupporting function calls

OS Spring’04

Environment External entities

TerminalOpen filesCommunication channels Local With other machines

OS Spring’04

Process control block (PCB)

statememory

files

accounting

priorityuser

CPU registersstorage

text

data

heap

stack

PSW

IR

PC

SP

generalpurposeregisters

PCB CPUkernel user

OS Spring’04

Process States

running

ready blocked

created

schedule

preempt

event done

wait for event

terminated

OS Spring’04

UNIX Process Statesrunning

user

runningkernel

readyuser

readykernel blocked

zombie

sys. callinterrupt

schedule

created

return

terminated

wait for event

event done

schedule

preempt

interrupt

OS Spring’04

Multiprocesses mechanisms Context switch Create process and dispatch (in

Unix, fork()/exec()) End process (in Unix, exit() and

wait())

OS Spring’04

Threads Thread: an execution within a process A multithreaded process consists of

many co-existing executions Separate:

CPU state, stack

Shared:Everything else Text, data, heap, environment

OS Spring’04

Thread Support Operating system

Advantage: thread scheduling done by OS Better CPU utilization

Disadvantage: overhead if many threads

User-levelAdvantage: low overheadDisadvantage: not known to OS E.g., a thread blocked on I/O blocks all the

other threads within the same process

OS Spring’04

Multiprogramming Multiprogramming: having multiple

jobs (processes) in the systemInterleaved (time sliced) on a single CPUConcurrently executed on multiple CPUsBoth of the above

Why multiprogramming?Responsiveness, utilization, concurrency

Why not?Overhead, complexity

OS Spring’04

ResponsivenessJob 1arrives

Job 1terminates

Job1 Job2 Job3

Job 2terminates

Job 3terminates

Job 2arrives

Job 3arrives

Job1Job3

Job2

Job 1 terminates Job 3 terminates

Job 2 terminates

OS Spring’04

Workload matters!? Would CPU sharing improve

responsiveness if all jobs were taking the same time?

No. It makes it worse! For a given workload, the answer depends

on the value of coefficient of variation (CV) of the distribution of job runtimes

CV=stand. dev. / meanCV < 1 => CPU sharing does not helpCV > 1 => CPU sharing does help

OS Spring’04

Real workloads Exp. Dist: CV=1;Heavy Tailed Dist:

CV>1 Dist. of job runtimes in real systems is

heavy tailedCV ranges from 3 to 70

Conclusion: CPU sharing does improve

responsiveness CPU sharing is approximated byTime slicing: interleaved execution

OS Spring’04

Utilization

idle

idle

idle

idle

idle

idle

1st I/Ooperation

I/Oends

2nd I/Ooperation

I/Oends

3rd I/Ooperation

CPU

Disk

CPU

Disk idle idle

idle

idleJob1 Job1

Job1 Job1Job2

Job2

Job2

OS Spring’04

Workload matters!? Does it really matter?Yes, of course:

If all jobs are CPU bound (I/O bound), multiprogramming does not help to improve utilization

A suitable job mix is created by a long-term scheduling

Jobs are classified on-line to be CPU (I/O) bound according to the job’s history

OS Spring’04

Concurrency Concurrent programming

Several process interact to work on the same problem ls –l | more

Simultaneous execution of related applications Word + Excel + PowerPoint

Background execution Polling/receiving Email while working on

smth else

OS Spring’04

The cost of multiprogramming Switching overhead

Saving/restoring context wastes CPU cycles

Degrades performanceResource contentionCache misses

ComplexitySynchronization, concurrency control, deadlock avoidance/prevention

OS Spring’04

Short-Term Scheduling

running

ready blocked

created

schedule

preempt

event done

wait for event

terminated

OS Spring’04

Short-Term scheduling Process execution pattern consists

of alternating CPU cycle and I/O wait CPU burst – I/O burst – CPU burst – I/O

burst...

Processes ready for execution are hold in a ready (run) queue

STS schedules process from the ready queue once CPU becomes idle

OS Spring’04

Metrics: Response time

Job arrives/becomes ready to run Starts running

Job terminates/blocks waiting for I/O

TwaitTrun

Tresp

Tresp= Twait + Trun

Response time (turnaround time) is the average over the jobs’Tresp

OS Spring’04

Other Metrics

Wait time: average of Twait

This parameter is under the system control

Response ratio or slowdownslowdown=Tresp / Trun

Throughput, utilization depend on user imposed workload=>

Less useful

OS Spring’04

Note about running time (Trun)

Length of the CPU burstWhen a process requests I/O it is still “running” in the systemBut it is not a part of the STS workload

STS view: I/O bound processes are short processes

Although text editor session may last hours!

OS Spring’04

Off-line vs. On-line scheduling Off-line algorithms

Get all the information about all the jobs to schedule as their inputOutputs the scheduling sequencePreemption is never needed

On-line algorithmsJobs arrive at unpredictable timesVery little info is available in advancePreemption compensates for lack of knowledge

OS Spring’04

First-Come-First-Serve (FCFS) Schedules the jobs in the order in

which they arriveOff-line FCFS schedules in the order the jobs appear in the input

Runs each job to completion Both on-line and off-line Simple, a base case for analysis Poor response time

OS Spring’04

Shortest Job First (SJF) Best response time

Short Long job

Long job Short

Inherently off-lineAll the jobs and their run-times must be available in advance

OS Spring’04

Preemption Preemption is the action of stopping

a running job and scheduling another in its place

Context switch: Switching from one job to another

OS Spring’04

Using preemption On-line short-term scheduling algorithms

Adapting to changing conditions e.g., new jobs arrive

Compensating for lack of knowledge e.g., job run-time

Periodic preemption keeps system in control

Improves fairnessGives I/O bound processes chance to run

OS Spring’04

Shortest Remaining Time first (SRT)

Job run-times are known Job arrival times are not known When a new job arrives:

if its run-time is shorter than the remaining time of the currently executing job:preempt the currently executing job and schedule the newly arrived job

else, continue the current job and insert the new job into a sorted queue

When a job terminates, select the job at the queue head for execution

OS Spring’04

Round Robin (RR) Both job arrival times and job run-

times are not known Run each job cyclically for a short

time quantumApproximates CPU sharing

Job 1arrives

Job1 3

Job 2arrives

Job 3arrives

2 1 2 31 2 31 2 1 Job2

Job 3terminates

Job 1terminates

Job 2terminates

OS Spring’04

ResponsivenessJob 1arrives

Job 1terminates

Job1 Job2 Job3

Job 2terminates

Job 3terminates

Job 2arrives

Job 3arrives

Job1Job3

Job2

Job 1 terminates Job 3 terminates

Job 2 terminates

OS Spring’04

Priority Scheduling RR is oblivious to the process past

I/O bound processes are treated equally with the CPU bound processes

Solution: prioritize processes according to their past CPU usage

16

1

8

1

4

1

2

1:

2

1

10 ,)1(

3211

1

nnnnn

nnn

TTTTE

ETE

Tn is the duration of the n-th CPU burst En+1 is the estimate of the next CPU burst

OS Spring’04

Multilevel feedback queues

quantum=10

quantum=20

quantum=40

FCFS

new jobs terminated

OS Spring’04

Multilevel feedback queues Priorities are implicit in this scheme Very flexible Starvation is possibleShort jobs keep arriving => long jobs

get starved Solutions:

Let it beAging

OS Spring’04

Priority scheduling in UNIX Multilevel feedback queues

The same quantum at each queueA queue per priority

Priority is based on past CPU usagepri=cpu_use+base+nice

cpu_use is dynamically adjustedIncremented each clock interrupt: 100 sec-1

Halved for all processes: 1 sec-1

OS Spring’04

Fair Share scheduling Given a set of processes with

associated weights, a fair share scheduler should allocate CPU to each process in proportion to its respective weight

Achieving pre-defined goalsAdministrative considerations Paying for machine usage, importance of the

project, personal importance, etc.

Quality-of-service, soft real-time Video, audio

OS Spring’04

Perfect Fairness

),( interval time theduringA processby

received CPU ofamount :),(

Aclient of share alproportion :

)(),(

21

21

1221

tt

ttW

S

S

SttttW

A

A

i i

AA

A fair share scheduling algorithm achieves perfect fairness in time interval (t1,t2) if

OS Spring’04

Wellness criterion for FSS

• An ideal fair share scheduling algorithm achieves perfect fairness for all time intervals

The goal of a FSS algorithm is to receive CPU allocations as close as possible to a perfect FSS algorithm

OS Spring’04

Fair Share scheduling: algorithms

Weighted Round RobinShares are not uniformly spread in time

Lottery scheduling:Each process gets a number of lottery tickets proportional to its CPU allocationThe scheduler picks a ticket at random and gives it to the winning clientOnly statistically fair, high complexity

OS Spring’04

Fair Share scheduling: VTRR Virtual Time Round Robin (VTRR)

Order ready queue in the order of decreasing shares (the highest share at the head)Run Round Robin as usualOnce a process that exhausted its share is encountered:Go back to the head of the queue