CPU Scheduling

Presentation Transcript

CS423UG Operating Systems

Indranil Gupta

Lecture 5

Sep 2, 2005

Schedule for Today’s Lecture

Why Scheduling?
Scheduling Levels
Basic Scheduling Algorithm (FCFS)

CS 423UG - Operating Systems,Indranil Gupta

Oh, the Process!

What is a Process?

It’s one executing instance of a “program”
Consists of code, data, heap, stack, PC, regs, and a little other information in the PCB

Process state, priority, accounting
Program counter, register variables, stack pointers, etc
Open files and devices

It can start/manipulate/kill other processes

Threads: Lightweight Processes

A thread is a single execution

Stack
Program counter
Registers

All threads within a process share resources

Address space
Text, data, heap
Open files, etc

Implementations of threads

User-level, Kernel-level, Hybrid, Pop-up

Scheduling

Deciding which process/thread should occupy a resource (CPU, disk, etc.)

(CPU (horsepower))

I want to ride it

Whose turn is it?

Process 2

Process 3

Process 1

When to schedule?

“to schedule”=change the process that is in “running” state

Options for when to schedule:

A new process starts
The running process exits
The running process is blocked/makes a syscall
I/O interrupt received (some processes may now be ready)
Clock interrupt (every 10 milliseconds)

Preemptive vs. Non-preemptive

Non-preemptive scheduling:

The running process keeps the CPU until it voluntarily gives up the CPU

process exits
switches to blocked state
1and 4 only (no 3)

Preemptive scheduling:

The running process can be preempted and must release the CPU (can be forced to give up CPU)

4

Terminated

Running

1

3

Ready

Blocked

What are objectives of scheduling?

(CPU (horsepower))

I want to ride it

Whose turn is it?

Process 2

Process 3

Process 1

Scheduling Objectives

Fairness (nobody cries)
Priority (ladies first)
Efficiency (make best use of equipment)
Encourage good behavior (good boy/girl)
Support heavy loads (degrade gracefully)
Adapt to different environments (interactive, real-time, multimedia)

Concretely: Performance Metrics

Efficiency: percentage of time a resource is being used
Throughput: # of processes that complete per unit time
Turnaround Time (also called elapse time)

(process finish time – process entry time) : could be > process’s own total running time (why?)

Waiting Time

total amount of time process waits (in ready state)

Response Time

amount of time from when a request was first submitted until first response is produced: useful for interactive processes

Others

Fairness
Policy Enforcement: seeing that stated policy is carried out
Proportionality: meet users' expectation
Meeting Deadlines: avoid losing data

Different Systems, Different Focuses

For all

Fairness, policy enforcement, resource efficiency, resource balance

Batch Systems (e.g., supercomputers)

Max throughput, min turnaround time, max CPU utilization

Interactive Systems (e.g., PCs)

Min Response time, best proportionality

Real-Time Systems (e.g., Space Shuttle)

Predictability, meeting deadlines

Process Profiles

CPU – Bound

Time spent in I/O < Time spent computing in CPU

I/O – Bound

Time spent in I/O > Time spent computing in CPU

Process may spend time waiting for interrupts

Process Mix

Scheduling should balance load between I/O bound and CPU-bound processes
Ideal would be to run all equipment at 100% utilization but that would not necessarily be good for response time

Program Behaviors Considered in Scheduling

I/O

Is it I/O bound? Example?
Is it CPU bound? Example?
Batch or interactive environment
Urgency
Priority
Frequency of preemption
Frequency of page faults
How much execution time it has already received
How much execution time it needs to complete

I/O

compute

Scheduling Levels

“Job”=Process

Long-term scheduler: Which jobs allowed to compete for the CPU and other resources
Medium-term scheduler: Which jobs to temporarily suspend or resume to smooth fluctuations in system load
Short-term scheduler: Assigning CPU to a ready process

P3 P5 P8 P10 P2

Ready Queue

(of processes)

CPU Scheduler

Proc 1: 14 time units
Proc2: 8 time units
Proc3: 8 time units
Dispatcher

CPU

Dispatcher

Proc 1

Proc 10

Proc 2

Proc 11

Proc 3

Proc 12

Ready queue

Blocked queue

Dispatcher

Gives the control of the CPU to the process that is awarded CPU by short-term scheduler
Functions:

switching context
switching back to user mode
jumping to the proper location in the user process

Dispatch Latency = delay of above operations

Needs to be fast

Single Processor Scheduling Algorithms

Batch systems

First Come First Serve (FCFS)
Shortest Job First

Interactive Systems

Round Robin
Priority Scheduling
Multi Queue & Multi-level Feedback
Shortest process time
Guaranteed Scheduling
Lottery Scheduling
Fair Sharing Scheduling

First Come First Serve (FCFS)

Process that requests the CPU FIRST is allocated the CPU FIRST
Called “FIFO”, Non-preemptive, Used

in Batch Systems

Real life analogy: Any ticket counter
Implementation: FIFO queues

A new process enters the tail of the queue
The scheduler selects from the head of the queue.

Performance Metric: Average Waiting Time (AWT)
Given Parameters:

Burst Time (in ms), Arrival Time and Order
Can be generalized to processes with alternate CPU and I/O bursts: blocking process goes to queue’s tail

FCFS Example

The final schedule:

P1 (24)

P2 (3)

P3 (4)

24

27

0

P1 waiting time: 0-0

The average waiting time: (0+21+23)/3 = 14.667

Problems with FCFS

Non-preemptive
Not optimal average waiting time (AWT)
Cannot utilize resources in parallel:

Suppose there is 1 CPU-bound process and many I/O-bound processes
Result: Convoy effect, low CPU and I/O Device utilization

What are Convoy Effects?

Consider n-1 jobs in system that are I/O bound and 1 job that is CPU bound.
I/O bound jobs pass quickly through the ready queue and suspend themselves waiting for I/O.
CPU bound job arrives at head of queue and executes until complete.
I/O bound jobs rejoin ready queue and wait for CPU bound job to complete.
I/O devices idle until CPU bound job completes.
When CPU bound job completes, other processes rush to wait on I/O again.
CPU becomes idle.