Вы находитесь на странице: 1из 32

Operating System Examples Scheduling


er/ch10.html http://linuxbangalore.org/blug/meetings/200401/scheduler2.6.pdf http://josh.trancesoftware.com/linux/linux_cpu_s cheduler.pdf http://www.linuxinsight.com/files/sched-designCFS.txt Windows Operating System Internals Curriculum Development Kit, developed by Solomon, Russinovich and Polze

Solaris uses priority- Default scheduling

based thread class scheduling time sharing Each thread/process Within each class: belongs to one of six Different priorities classes Different scheduling
Time sharing Interactive Real time System File share Fixed priority


Time sharing scheduling algorithm

alters priorities of processes Uses a multilevel feedback queue The higher the priority the smaller the time slice The lower the priority the larger the time slice Interactive processes typically have higher priority (good response time) CPU-bound processes have lower priority (good throughput)

Time sharing scheduling algorithm

a thread has used its entire time quantum without blocking its priority is changed The priority of a thread that is returning from sleeping (e.g. waiting for I/O) has its priority increased
Interactive scheduling algorithm Similar to time sharing but give window managers (e.g., KDE, GNOME) a higher priority

Solaris Dispatch Table

Solaris Dispatch Table

Priority: A higher number indicates a

higher priority Time quantum: There is an inverse relationship between priorities and time quanta Time quantum expired: The new priority of a thread that has used its entire time quantum without blocking Return from sleep: The priority of a thread that is returning from sleeping (such as waiting for I/O)

Solaris has 7 categories

Real-time class Highest priority A thread in this class will before a process in any other class Few processes belong in this class

System class Kernel threads are assigned to this class e.g., scheduler and paging daemons


in Solaris 9 Same number of priorities as time sharing but no dynamic adjustment of priorities
Fair-share Uses CPU shares instead of priorities A share is an indicator of entitlement

Each scheduling class includes a set of

priorities However, the scheduler converts the classspecific priorities into global priorities The thread with the highest global priority is chosen to run

Solaris Scheduling

Windows XP
The scheduler is called a dispatcher 32-level priority scheme Priorities are divided into two classes: Variable class: priorities 1 to 15 Real-time class: priorities 16 to 31 There is a queue for each priority The dispatcher traverses the set of

queues from highest to lowest until it finds a thread that is ready to run If there are no threads ready to run the dispatcher executes the idle thread

Windows XP Scheduling
Processes are given a priority class upon



Priorities in all classes (except the

REALTIME_PRIORITY) class can change

Windows XP Scheduling
A thread within a given priority class has a

relative priority within the class:


Windows XP
Each thread has a base priority that

represents a value in the priority range for the class that the thread belongs The priority of each thread is based on both the priority class it belongs to and its relative priority within that class As with Solaris the classes and the priorities in the classes are mapped to a global priority

Windows XP Priorities

Windows XP
Priority changes for threads in the variable


Time quantum expires:

If thread is in the variable class the priority is lowered but not below the base priority

Thread switches from blocked to running:

Priority is increased The amount depends on what the thread was doing Keyboard I/O gets a large increase while disk I/O gets a moderate increase

Windows XP
Windows XP distinguishes between the

foreground process that is currently selected on the screen and the background processes that are not currently selected The strategy tends to give good response times to interactive threads that are using the mouse and windows The I/O devices are kept busy

Linux Scheduling

fully O(1) scheduling. Implement SMP scalability. Provide good interactive performance. Provide fairness. Optimize for the common case of only one or two runnable processes, yet scale well to multiple processors, each with many processes.

Linux Scheduling
The Linux kernel internally represents

processes and threads as tasks via the task_struct Linux uses task_struct (process control block)to represent any execution context
A single-threaded process will be represented with one task structure A multithreaded process will have one task structure for each of the user-level threads

Linux Scheduling
Preemptive, priority-based

Constant order O(1) scheduling time

Two priority ranges: time-sharing and real-

time Real-time range from 0 to 99 and nice value from 100 to 140 Linux (unlike Solaris) assigns higher priority tasks longer time quanta and lower priority tasks shorter time quanta

Linux Scheduling

Priorities and Time-slice length

Linux Scheduling
A runnable task is considered for

execution on the CPU as long as it has time remaining in its time slice When a task has exhausted its time slice, it is considered expired and is not eligible for execution again until all other tasks have all exhausted their time slice

Linux Scheduling
The kernel maintains a list of all runnable

tasks in a runqueue data structure In a multi-core machine there is a runqueue for each processor Each runqueue consists of two priority arrays:
Active: Contains all tasks with time remaining in their time slices Expired: Contains all expired tasks

The active, expired queues are indexed

according to priority

Linux Scheduling

List of Tasks Indexed According to Priorities

Linux Scheduling
The scheduler chooses the task with the

highest priority from the active array for execution When there are no more tasks in the active array, the expired array becomes the active array and the active array becomes the expired array.

Linux Scheduling
The scheduler chooses the task with the

highest priority from the active array for execution A tasks dynamic priority is recalculated when the task has exhausted its time quantum and is to be moved to the expired array

Linux Scheduling
Linux tries to give a higher

interactive tasks The scheduler somehow needs to measure the interactivity of a process Interactive processes sleep more than others

priority to

Linux Scheduling
The scheduler keeps track of the average

amount of time a task sleeps vs average amount of time task uses CPU

Use the sleep_avg variable

When a process wakes the time spent

sleeping is added to its sleep_avg; The run time is subtracted from sleep_avg The higher the value of sleep_avg the more I/O bound it is


Can Linux processes starve?

Case Study Discussion

Basic idea for schedule user processes is the same

for all systems:

Lower priority for CPU bound process Increase priority for I/O bound process

The scheduling in Solaris / Linux is more

concerned about fairness.

More popular as the OSes for servers.

The scheduling in Window XP is more concerned

about user perceived performance.

More popular as the OS for personal computers.

We have examined scheduling in several

contemporary operating systems