Chapter 5: CPU Scheduling. Operating System Concepts 8 th Edition,

Transcription

1 Chapter 5: CPU Scheduling, Silberschatz, Galvin and Gagne 2009

2 Chapter 5: CPU Scheduling Basic Concepts Scheduling Criteria Scheduling Algorithms Thread Scheduling Multiple-Processor Scheduling Linux Example 5.2 Silberschatz, Galvin and Gagne 2009

3 Objectives To introduce CPU scheduling, which is the basis for multiprogrammed operating systems To describe various CPU-scheduling algorithms To discuss evaluation criteria for selecting a CPU-scheduling algorithm for a particular system 5.3 Silberschatz, Galvin and Gagne 2009

4 Basic Concepts The objective of multiprogramming is to have some process running at all time, to Maximum CPU utilization. CPU I/O Burst Cycle Process execution consists of a cycle of CPU execution and I/O wait Process execution begins with a CPU burst that is followed by an I/O burst, which is followed by another CPU burst, then another I/O burst, and so on,.. The final CPU burst ends the process. CPU burst distribution large number of short CPU bursts and a small number of long CPU bursts. An I/O bound program has many short CPU bursts. A CPU bound program has few long CPU bursts. 5.4 Silberschatz, Galvin and Gagne 2009

5 Histogram of CPU-burst Times 5.5 Silberschatz, Galvin and Gagne 2009

6 Alternating Sequence of CPU And I/O Bursts 5.6 Silberschatz, Galvin and Gagne 2009

7 CPU Scheduler When the CPU becomes idle, the OS must Select from among the processes in memory that are ready to execute, and allocates the CPU to one of them. The selection process is carried out by the short-term scheduler (CPU scheduler ). CPU scheduling decisions may take place when a process: 1. Switches from running state to the waiting state(result of I/o request or wait for the termination of one of the child processes). 2. Switches from running state to ready state(interrupt). 3. Switches from waiting state to ready state(completion of I/O) 4. Terminates Scheduling under 1 and 4 is nonpreemptive or cooperative. All other scheduling is preemptive 5.7 Silberschatz, Galvin and Gagne 2009

8 Preemptive scheduling Under nonpreemptive scheduling, once the CPU has been allocated to a process, the process keeps the CPU until it releases the CPU either by terminating or by switching to the waiting state. Windows 95 and all subsequent versions of windows OS have used preemptive scheduling. 5.8 Silberschatz, Galvin and Gagne 2009

9 Dispatcher The Dispatcher is the module that 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 It should be fast. Dispatch latency the time it takes for the dispatcher to stop one process and start another running 5.9 Silberschatz, Galvin and Gagne 2009

10 Scheduling Criteria CPU utilization keep the CPU as busy as possible. Throughput # of processes that complete their execution per time unit(10 processes/second) Turnaround time amount of time to execute a particular process(the interval from the time of submission of a process to the time of completion, waiting to get into memory, waiting in the ready queue, exciting on the CPU, doing I/O). Waiting time the amount of times 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 output (for timesharing environment) 5.10 Silberschatz, Galvin and Gagne 2009

11 Scheduling Algorithm Optimization Criteria Max CPU utilization Max throughput Min turnaround time Min waiting time Min response time 5.11 Silberschatz, Galvin and Gagne 2009

12 First-Come, First-Served (FCFS) Scheduling Jobs are scheduled in order of arrival When a process enters the ready queue, its PCB is linked onto the tail of the queue. When the CPU is free, it is allocated to the process at the head of the queue (the running process is then removed from the queue). Disadvantages: Non-preemptive : once the CPU is allocated to a process, the process keeps the CPU until it releases it, either by terminating or requesting I/O. The average waiting time is often quite long Silberschatz, Galvin and Gagne 2009

13 example Process Burst 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: P 1 P 2 P Waiting time for P 1 = 0; P 2 = 24; P 3 = 27 Average waiting time: ( )/3 = Silberschatz, Galvin and Gagne 2009

14 FCFS Scheduling (Cont) Suppose that the processes arrive in the order P 2, P 3, P 1 The Gantt chart for the schedule is: P 2 P 3 P 1 0 Waiting time for P 1 = 6; P 2 = 0 ; P 3 = 3 Average waiting time: ( )/3 = 3 Much better than previous case Convoy effect as short processes go behind long process lower CPU and device utilization Silberschatz, Galvin and Gagne 2009

15 Shortest-Job-First (SJF) Scheduling This algorithm Associates with each process the length of its next CPU burst. Use these lengths to schedule the process with the shortest time, if the next CPU bursts of two processes are the same, FCFS scheduling is used. Two schemes: Nonpreemptive once CPU given to the process it cannot be preempted until 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 the Shortest-Remaining-Time-First (SRTF) 5.15 Silberschatz, Galvin and Gagne 2009

16 Examples of SJF Example1: Process Arrival Time Burst Time P P P P SJF scheduling chart P 4 P P 3 1 P Average waiting time = ( ) / 4 = 7 Compare with FCFS AWT=( )/4= Silberschatz, Galvin and Gagne 2009

17 Shortest-Job-First (SJF) Scheduling Example2: Process Arrival Time Burst Time P P P P Non preemptive SJF Average waiting time = ( )/4 = 4 P 1 P 3 P 2 P P 1 (7) P 2 (4) P 3 (1) P 4 (4) P 1 s wating time = 0 P 2 s wating time = 6 P 3 s wating time = 3 P 4 s wating time = Silberschatz, Galvin and Gagne 2009

18 Shortest-Job-First (SJF) Scheduling Example3: Process Arrival Time Burst Time P P P P Preemptive SJF(SRTF) Average waiting time = ( )/4 = 3 P 1 P 2 P 3 P 2 P 4 P P P 1 (7) 1 (5) P 2 (4) P 2 (2) P 3 (1) P 4 (4) 16 P 1 s wating time = 9 P 2 s wating time = 1 P 3 s wating time = 0 P 4 s wating time = Silberschatz, Galvin and Gagne 2009

19 Shortest-Job-First (SJF) Scheduling SJF is optimal gives minimum average waiting time for a given set of processes The difficulty is knowing the length of the next CPU request Silberschatz, Galvin and Gagne 2009

20 Prediction of the Length of the Next CPU Burst 5.20 Silberschatz, Galvin and Gagne 2009

21 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 in Unix but lowest in Java) Equal-priority processes are scheduled in FCFS order. Preemptive: preempt the CPU if the priority of the newly arrived process is higher than the priority of the currently running process. Nonpreemptive : put the new process at the head of the ready queue. 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 (for example : 1 every 15 minutes) 5.21 Silberschatz, Galvin and Gagne 2009

22 Priority Scheduling Example : Process Burst Time priority P P P P p5 5 2 All arrived at time 0. The Gantt chart for the schedule is: P 2 P P 1 5 P 3 P The AWT is ( )/5 = Silberschatz, Galvin and Gagne 2009

23 Priority Scheduling Example: Process arrival time Burst length Priority P P P P P Gantt chart: Non-preemptive priority scheduling P2 P1 P5 P3 P Gantt chart: Preemptive priority scheduling P2 P1 P5 P1 P3 P Silberschatz, Galvin and Gagne 2009

24 Round Robin (RR) Is designed especially for time-sharing systems. Similar to FCFS, but it is Preemptive to enable the system to switch between processes. Each process gets a small unit of CPU time (time quantum or time slice), usually milliseconds. The Ready queue is FIFO (new processes are added to the tail of the queue.) The CPU scheduler picks the first process from the ready queue,set a timer to interrupt after 1 time quantum, and dispatch the process Silberschatz, Galvin and Gagne 2009

25 Round Robin (RR) One of two things will happen The process may have a CPU burst of < 1 time quantum the process itself will release the CPU voluntarily. The CPU burst of the currently running process > 1 time quantum the timer will go off and will cause an interrupt to the OS. a context switch will be executed, and the process will be put at the tail of the ready queue. The CPU scheduler will then select the next process in the ready queue. Typically, higher average turnaround than SJF, but better response 5.25 Silberschatz, Galvin and Gagne 2009

26 Round Robin (RR) Example1: Time quantum = 4 Process Burst Time P 1 24 P 2 3 P 3 3 The Gantt chart is: P 1 P 2 P 3 P 1 P 1 P 1 P 1 P AWT(6(10-4)+4+7)/3 = Silberschatz, Galvin and Gagne 2009

27 Example2: Round Robin (RR) Time quantum = 20 Process Burst Time Wait Time P = 81 P P = 94 P = 97 P 1 P 2 P 3 P 4 P 1 P 3 P 4 P 1 P 3 P P 1 (53) P 1 (33) 24 P 1 (13) P 2 (17) P 3 (68) P 4 (24) P 3 (48) P 3 (28) P 3 (8) 40 P 4 (4) 17 Average wait time = ( )/4 = Silberschatz, Galvin and Gagne 2009

28 Round Robin (RR) 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 until its next time quantum. (Ex: 5 processes, TQ = 20 milliseconds, each process will get up to 20 milliseconds every 100 milliseconds. The Performance of RR depends heavily on the size of the TQ. TQ large FCFS TQsmall TQ must be large (but not too large)with respect to context switch time, otherwise overhead is too high 5.28 Silberschatz, Galvin and Gagne 2009

29 Time Quantum and Context Switch Time 5.29 Silberschatz, Galvin and Gagne 2009

30 Turnaround Time Varies With The Time Quantum The average TurnAroundTime of a set of process does not necessarily improve as the TQ size increase. The AVG TAT can be improved if most process finish their next CPU burst in a single time quantum Silberschatz, Galvin and Gagne 2009

31 Multilevel Queue Processes are classified into different groups. Each group have different response-time requirements different scheduling needs. A multilevel queue scheduling algorithm partitions the Ready queue 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 preemptive 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 5.31 Silberschatz, Galvin and Gagne 2009

32 Multilevel Queue Scheduling 5.32 Silberschatz, Galvin and Gagne 2009

33 Multilevel Feedback Queue Implement multiple ready queues Different queues may be scheduled using different algorithms Just like multilevel queue scheduling, but assignments are not static Multilevel feedback queue-scheduling algorithm allows a process to 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 and downgrade a process The most general CPU-scheduling algorithm. The most complex algorithm Silberschatz, Galvin and Gagne 2009

34 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. AT Q 2 job is served FCFS only when queue 0 and queue 1 are empty Silberschatz, Galvin and Gagne 2009

35 Multilevel Feedback Queues 5.35 Silberschatz, Galvin and Gagne 2009

36 Thread Scheduling Distinction between user-level and kernel-level threads On OSs that support them, it is the kernel-level threadsnot processes- that are being scheduled by OS. User-level threads are managed by a thread library and the kernel is unaware of them. To run on CPU, the user level threads must be mapped to an associated kernel-level thread. It may use a lightweight process(lwp). contention scope: one distinction between user-level and kernel-level threads lies in how they are scheduled Silberschatz, Galvin and Gagne 2009

37 Thread Scheduling Many-to-one and many-to-many models, thread library schedules user-level threads to run on LWP. Known as process-contention scope (PCS) since scheduling competition takes place among threads belonging to the same process. PCS is done according to preempt priority. PTHREAD SCOPE PROCESS schedules threads using PCS scheduling. Kernel thread scheduled onto available CPU is system-contention scope (SCS) competition takes place among all threads in system Systems using the one-to-one model schedule threads using only SCS. PTHREAD SCOPE SYSTEM schedules threads using SCS scheduling Silberschatz, Galvin and Gagne 2009

38 Multiple-Processor Scheduling CPU scheduling more complex when multiple CPUs are available Different rules for homogeneous processors (Identical processors in terms of their functionality) or heterogeneous processors. Asymmetric multiprocessing: All scheduling decisions, I/O processing, and other system activities handled by a single processor the master server. The other processors execute only user code. Simple because only one processor accesses the system data structures, reducing the need for data sharing. Symmetric multiprocessing (SMP): each processor is self-scheduling, all processes in common ready queue, or each has its own private queue of ready processes Multiple processors try to access and update a common data structures. So, scheduler must be programmed carefully. Must ensure that 2 processors don t choose the same process Silberschatz, Galvin and Gagne 2009

39 Linux Scheduling Linux Scheduler is a preemptive, priority-based algorithm with 2 separate priority ranges: Two priority ranges: time-sharing and real-time A real-time range from 0 to 99 Longer time quantum A nice value ranging from 100 to 140 Shorter time quantum 5.39 Silberschatz, Galvin and Gagne 2009

40 Linux Scheduling The kernel maintains a list of all runnable tasks in a runqueue data structure. Each runqueue contains two priority arrays : Active : contains all tasks with time remaining in their time slices Expired : contains all expired tasks Silberschatz, Galvin and Gagne 2009

41 List of Tasks Indexed According to Priorities The scheduler chooses the task with the highest priority from the active array for execution on the CPU. When the active array is empty the 2 arrays are exchanged (the expired array becomes the active array, and vice versa) Silberschatz, Galvin and Gagne 2009

42 Algorithm Evaluation Deterministic modeling takes a particular predetermined workload and defines the performance of each algorithm for that workload More examples P: 214 Simple and fast Requires exact numbers for input and its answers apply only for those data Queueing models rate at which new processes arrive, ratio of CPU bursts to I/O times, distribution of CPU burst times and I/O burst times can be measured and then approximated or estimated result is a mathematical formula describing it From these it is possible to compute the average throughput, utilization, waiting time, and so on difficult to work 5.42 Silberschatz, Galvin and Gagne 2009

43 Algorithm Evaluation Simulations run computer simulations of the different proposed algorithms data to drive the simulation can be randomly generated better alternative when possible is to generate trace tapes expensive Implementation The only completely accurate way to evaluate a scheduling algorithm is to code it up, put it in the operating system, and see how it works. high cost (coding and user reaction) 5.43 Silberschatz, Galvin and Gagne 2009

44 Simulations 5.44 Silberschatz, Galvin and Gagne 2009

45 Conclusion We ve looked at a number of different scheduling algorithms. Which one works the best is application dependent. General purpose OS will use priority based, round robin, preemptive Real Time OS will use priority, no preemption Silberschatz, Galvin and Gagne 2009

46 End of Chapter 5, Silberschatz, Galvin and Gagne 2009