Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Silberschatz, Galvin and Gagne ©2007 Chapter 5: CPU Scheduling

5.2 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Chapter 5: CPU Scheduling  Basic Concepts  Scheduling Criteria  Scheduling Algorithms  Multiple-Processor Scheduling  Real-Time Scheduling  Thread Scheduling  Operating Systems Examples  Algorithm Evaluation

5.3 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Basic Concepts  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 distribution

5.4 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Alternating Sequence of CPU And I/O Bursts

5.5 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Histogram of CPU-burst Times

5.6 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 CPU Scheduler  Selects from among the processes in memory that are ready to execute, and allocates the CPU to one of them  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  Scheduling under 1 and 4 is nonpreemptive  All other scheduling is preemptive

5.7 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Dispatcher  Dispatcher module gives control of the CPU to the process selected by the short-term 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

5.8 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Scheduling Criteria  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, not total output (for time-sharing environment) E.g., when a Web page starts rendering (response time) vs. when it finishes (turnaround time)  Fairness – does every job get a chance to run

5.9 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Optimization Criteria  Max CPU utilization: percentage of time CPU is not idle  Max throughput: processes completed over time  Min turnaround time: submission to completion  Min waiting time: time spent waiting in the queue for service  Min response time: submission to beginning of response

5.10 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 First-Come, First-Served (FCFS) Scheduling ProcessBurst Time P 1 24 P 2 3 P 3 3  Suppose that the processes arrive in the order: P 1, P 2, P 3 The Gantt Chart for the schedule is:  Waiting time for P 1 = 0; P 2 = 24; P 3 = 27  Average waiting time: ( )/3 = 17 P1P1 P2P2 P3P

5.11 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 FCFS Scheduling (Cont.) Suppose that the processes arrive in the order P 2, P 3, P 1  The Gantt chart for the schedule is:  Waiting time for P 1 = 6; P 2 = 0 ; P 3 = 3  Average waiting time: ( )/3 = 3  Much better than previous case  Convoy effect short process behind long process Also called head of the line blocking P1P1 P3P3 P2P

5.12 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Shortest-Job-First (SJF) Scheduling  Associate with each process the length of its next CPU burst. Use these lengths to schedule the process with the shortest time  Two schemes: nonpreemptive – once CPU given to the process it cannot be preempted until process completes its CPU burst preemptive – if a new process arrives with CPU burst length less than remaining time of current executing process, preempt. This scheme is known as Shortest-Remaining-Time-First (SRTF)  SJF is optimal – gives minimum average waiting time for a given set of processes

5.13 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 ProcessArrival TimeBurst Time P P P P  SJF (non-preemptive)  Average waiting time = ( )/4 = 4 Example of Non-Preemptive SJF P1P1 P3P3 P2P P4P4 812

5.14 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Example of Preemptive SJF ProcessArrival TimeBurst Time P P P P  SJF (preemptive)  Average waiting time = ( )/4 = 3 P1P1 P3P3 P2P P4P4 57 P2P2 P1P1 16

5.15 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Determining Length of Next CPU Burst  Can only estimate the length  Can be done by using the length of previous CPU bursts, using exponential averaging

5.16 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Prediction of the Length of the Next CPU Burst

5.17 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Examples of Exponential Averaging   =0  n+1 =  n Recent history does not count   =1  n+1 =  t n Only the actual last CPU burst counts  If we expand the formula, we get:  n+1 =  t n +(1 -  )  t n -1 + … +(1 -  ) j  t n -j + … +(1 -  ) n +1  0  Since both  and (1 -  ) are less than or equal to 1, each successive term has less weight than its predecessor

5.18 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Priority Scheduling  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 a priority scheduling where priority is the predicted next CPU burst time  Problem  Starvation – low priority processes may never execute  Solution  Aging – as time progresses increase the priority of the process

5.19 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Round Robin (RR)  Each process gets a small unit of CPU time (time quantum), usually 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.  Performance q large  FIFO q small  q must be large with respect to context switch, otherwise overhead is too high

5.20 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Example of RR with Time Quantum = 20 ProcessBurst Time P 1 53 P 2 17 P 3 68 P 4 24  The Gantt chart is:  Typically, higher average turnaround than SJF, but better response P1P1 P2P2 P3P3 P4P4 P1P1 P3P3 P4P4 P1P1 P3P3 P3P

5.21 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Time Quantum and Context Switch Time

5.22 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Turnaround Time Varies With The Time Quantum

5.23 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Multilevel Queue  Ready queue is partitioned into separate queues: foreground (interactive) background (batch)  Each queue has its own scheduling algorithm foreground – RR background – FCFS  Scheduling must be done between the queues Fixed priority scheduling; (i.e., serve all from foreground then from background). Possibility of starvation. Time slice – each queue gets a certain amount of CPU time which it can schedule amongst its processes; i.e., 80% to foreground in RR 20% to background in FCFS

5.24 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Multilevel Queue Scheduling

5.25 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Multilevel Feedback Queue  A process can move between the various queues; aging can be implemented this way  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

5.26 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Example of Multilevel Feedback Queue  Three queues: Q 0 – RR with time quantum 8 milliseconds Q 1 – RR time quantum 16 milliseconds Q 2 – FCFS  Scheduling A new job enters queue Q 0 which is served FCFS. When it gains CPU, job receives 8 milliseconds. If it does not finish in 8 milliseconds, job is moved to queue Q 1. At Q 1 job is again served FCFS and receives 16 additional milliseconds. If it still does not complete, it is preempted and moved to queue Q 2.

5.27 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Multilevel Feedback Queues

5.28 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Multiple-Processor Scheduling  CPU scheduling more complex when multiple CPUs are available  Homogeneous processors within a multiprocessor  Load balancing – keep workload evenly distributed  Processor affinity – benefits of past history in a processor Why run a process on the same processor?  Asymmetric multiprocessing – only one processor accesses the system data structures, alleviating the need for data sharing

5.29 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Real-Time Scheduling  Real-time processes have timing constraints Expressed as deadlines or rate requirements  Hard real-time systems – required to complete a critical task within a guaranteed amount of time  Soft real-time computing – requires that critical processes receive priority over less fortunate ones  Common RT scheduling policies: Rate monotonic  Just one scalar priority related to the periodicity of the job  Priority = 1/rate  Static Earliest deadline first (EDF)  Dynamic but more complex  Priority = deadline  Both require admission control to provide guarantees

5.30 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Thread Scheduling  Local Scheduling – How the threads library decides which thread to put onto an available LWP  Global Scheduling – How the kernel decides which kernel thread to run next

5.31 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Linux Scheduling  Two algorithms: time-sharing and real-time  Time-sharing Prioritized credit-based – process with most credits is scheduled next Credit subtracted when timer interrupt occurs When credit = 0, another process chosen When all processes have credit = 0, recrediting occurs  Based on factors including priority and history  Real-time Soft real-time Posix.1b compliant – two classes  FCFS and RR  Highest priority process always runs first

5.32 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 The Relationship Between Priorities and Time-slice length

5.33 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 List of Tasks Indexed According to Priorities

5.34 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Policy versus Mechanism  Separate what is done vs. how it is done  Put some basic mechanism (or mechanisms) in the kernel  Permit user processes to set parameters that control scheduling  Simple example: nice command

5.35 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Algorithm Evaluation  First: What criteria are you optimizing for? CPU utilization, throughput, response time, etc.  Deterministic modeling – Gantt charts takes a particular workload Evaluate performance of each algorithm for that workload  Queuing models Analytic approach using math, distributions, algorithms, assumptions  Simulation Distribution-driven or trace-based  Implementation Expensive Also dependent on workload

5.36 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Algorithm Evaluation Evaluation of CPU schedulers by simulation

5.37 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Algorithm Evaluation

5.38 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 The Importance of Workloads  What kind of workload? Networking: Web pages, VoIP calls, P2P, IM, … Database: Transactions, decision support systems CPU: scientific, financial, software development  Obtaining workloads Privacy, legality, competitive advantages, etc.  How representative are workloads? University vs. Corporation vs. Government vs. Home user US vs. Asia vs. Europe  Variation in the workload Time-of-day effects Size of jobs can span multiple orders of magnitude

Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Silberschatz, Galvin and Gagne ©2007 End of Chapter 5

5.40 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Bursts of CPU usage alternate with periods of waiting for I/O. (a) A CPU-bound process. (b) An I/O-bound process. Scheduling – Process Behavior Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.41 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, Batch 2. Interactive 3. Real time Categories of Scheduling Algorithms Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.42 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Scheduling algorithm goals under different circumstances. Scheduling Algorithm Goals Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.43 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006  First-come first-served  Shortest job first  Shortest remaining Time next Scheduling in Batch Systems Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.44 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 An example of shortest job first scheduling. (a) Running four jobs in the original order. (b) Running them in shortest job first order. Shortest Job First Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.45 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006  Round-robin scheduling  Priority scheduling  Multiple queues  Shortest process next  Guaranteed scheduling  Lottery scheduling  Fair-share scheduling Scheduling in Interactive Systems Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.46 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Round-robin scheduling. (a) The list of runnable processes. (b) The list of runnable processes after B uses up its quantum. Round-Robin Scheduling Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.47 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 A scheduling algorithm with four priority classes. Priority Scheduling Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.48 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 (a) Possible scheduling of user-level threads with a 50-msec process quantum and threads that run 5 msec per CPU burst. Thread Scheduling (1) Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.49 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 (b) Possible scheduling of kernel-level threads with the same characteristics as (a). Thread Scheduling (2) Example from Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall, Inc. All rights reserved

5.50 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Java Scheduling  Loosely-defined scheduling policy. A thread runs until: 1. It’s time quantum expires 2. It blocks for I/O 3. It exits its run() method Some systems may support preemption

5.51 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Java Scheduling  Priorities - values range from 1-10  MIN_PRIORITY is 1  NORM_PRIORITY is 5  MAX_PRIORITY is 10

5.52 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Java Scheduling Changing priority using setPriority()

5.53 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Java Scheduling Relationship between Java and Win32 Priorities

5.54 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Java Scheduling Java thread scheduling on Solaris

5.55 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Pthread Scheduling API

5.56 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Solaris Scheduling

5.57 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Solaris Dispatch Table

5.58 Silberschatz, Galvin and Gagne ©2007 Operating System Concepts with Java – 7 th Edition, Nov 15, 2006 Windows XP Priorities