Lecture 5 CPU Scheduling Contents n Why CPU

  • Slides: 33
Download presentation
Lecture 5: CPU Scheduling

Lecture 5: CPU Scheduling

Contents n Why CPU Scheduling n Scheduling Criteria & Optimization n Basic Scheduling Approaches

Contents n Why CPU Scheduling n Scheduling Criteria & Optimization n Basic Scheduling Approaches n Priority Scheduling n Queuing and Queues Organization n Scheduling Examples in Real OS n Deadline Real-Time CPU Scheduling AE 4 B 33 OSS Lecture 5 / Page 2 Silberschatz, Galvin and Gagne © 2005

Basic Concepts n Maximum CPU utilization obtained with multiprogramming n CPU–I/O Burst Cycle –

Basic Concepts n Maximum CPU utilization obtained with multiprogramming n CPU–I/O Burst Cycle – Process execution consists of a cycle of CPU execution and I/O wait n CPU burst distribution AE 4 B 33 OSS Lecture 5 / Page 3 Silberschatz, Galvin and Gagne © 2005

CPU Scheduler n Selects from among the processes in memory that are ready to

CPU Scheduler n Selects from among the processes in memory that are ready to execute, and allocates the CPU to one of them n 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 n Scheduling under 1 and 4 is nonpreemptive n All other scheduling is preemptive AE 4 B 33 OSS Lecture 5 / Page 4 Silberschatz, Galvin and Gagne © 2005

Dispatcher n Dispatcher module gives control of the CPU to the process selected by

Dispatcher n Dispatcher module gives control of the CPU to the process selected by the short-term scheduler; this involves: switching context l switching to user mode l jumping to the proper location in the user program to restart that program l n Dispatch latency – time it takes for the dispatcher to stop one process and start another running – overhead AE 4 B 33 OSS Lecture 5 / Page 5 Silberschatz, Galvin and Gagne © 2005

Scheduling Criteria & Optimization n CPU utilization – keep the CPU as busy as

Scheduling Criteria & Optimization n CPU utilization – keep the CPU as busy as possible l Maximize CPU utilization n Throughput – # of processes that complete their execution per time unit l Maximize throughput n Turnaround time – amount of time to execute a particular process l Minimize turnaround time n Waiting time – amount of time a process has been waiting in the ready queue l Minimize waiting time n Response time – amount of time it takes from when a request was submitted until the first response is produced, not output (for time-sharing and interactive environment ) l AE 4 B 33 OSS Minimize response time Lecture 5 / Page 6 Silberschatz, Galvin and Gagne © 2005

First-Come, First-Served (FCFS) Scheduling Process Burst Time P 1 24 P 2 3 P

First-Come, First-Served (FCFS) Scheduling Process Burst Time P 1 24 P 2 3 P 3 3 n 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 0 24 P 3 27 30 n Waiting time for P 1 = 0; P 2 = 24; P 3 = 27 n Average waiting time: (0 + 24 + 27)/3 = 17 AE 4 B 33 OSS Lecture 5 / Page 7 Silberschatz, Galvin and Gagne © 2005

FCFS Scheduling (Cont. ) Suppose that the processes arrive in the order P 2

FCFS Scheduling (Cont. ) Suppose that the processes arrive in the order P 2 , P 3 , P 1 n The Gantt chart for the schedule is: P 2 0 P 3 3 P 1 6 30 n Waiting time for P 1 = 6; P 2 = 0; P 3 = 3 n Average waiting time: (6 + 0 + 3)/3 = 3 n Much better than previous case n Convoy effect short process behind long process AE 4 B 33 OSS Lecture 5 / Page 8 Silberschatz, Galvin and Gagne © 2005

Shortest-Job-First (SJF) Scheduling n Associate with each process the length of its next CPU

Shortest-Job-First (SJF) Scheduling n Associate with each process the length of its next CPU burst. Use these lengths to schedule the process with the shortest time n Two schemes: nonpreemptive – once CPU given to the process it cannot be preempted until completes its CPU burst l preemptive – if a new process arrives with CPU burst length less than remaining time of current executing process, preempt. This scheme is know as the Shortest-Remaining-Time-First (SRTF) l n SJF is optimal – gives minimum average waiting time for a given set of processes AE 4 B 33 OSS Lecture 5 / Page 9 Silberschatz, Galvin and Gagne © 2005

Example of Non-Preemptive SJF Process P 1 0. 0 P 2 2. 0 P

Example of Non-Preemptive SJF Process P 1 0. 0 P 2 2. 0 P 3 4. 0 P 4 5. 0 Arrival Time 7 4 1 4 Burst Time n SJF (non-preemptive) P 1 0 3 P 3 7 P 2 8 P 4 12 16 n Average waiting time = (0 + 6 + 3 + 7)/4 = 4 AE 4 B 33 OSS Lecture 5 / Page 10 Silberschatz, Galvin and Gagne © 2005

Example of Preemptive SJF Process P 1 0. 0 P 2 2. 0 P

Example of Preemptive SJF Process P 1 0. 0 P 2 2. 0 P 3 4. 0 P 4 5. 0 Arrival Time 7 4 1 4 Burst Time n SJF (preemptive) P 1 0 P 2 2 P 3 4 P 2 5 P 4 7 P 1 11 16 n Average waiting time = (9 + 1 + 0 +2)/4 = 3 AE 4 B 33 OSS Lecture 5 / Page 11 Silberschatz, Galvin and Gagne © 2005

Determining Length of Next CPU Burst n Can only estimate the length n Can

Determining Length of Next CPU Burst n Can only estimate the length n Can be done by using the length of previous CPU bursts, using exponential averaging AE 4 B 33 OSS Lecture 5 / Page 12 Silberschatz, Galvin and Gagne © 2005

Examples of Exponential Averaging n =0 n+1 = n l Recent history does not

Examples of Exponential Averaging n =0 n+1 = n l Recent history does not count l n =1 n+1 = tn l Only the actual last CPU burst counts l n If we expand the formula, we get: n+1 = tn+(1 - ) tn -1 + … +(1 - )j tn -j + … +(1 - )n +1 0 n Since both and (1 - ) are less than or equal to 1, each successive term has less weight than its predecessor AE 4 B 33 OSS Lecture 5 / Page 13 Silberschatz, Galvin and Gagne © 2005

Priority Scheduling n A priority number (integer) is associated with each process n The

Priority Scheduling n A priority number (integer) is associated with each process n The CPU is allocated to the process with the highest priority (smallest integer highest priority) Preemptive l Nonpreemptive l n SJF is a priority scheduling where priority is the predicted next CPU burst time n Problem Starvation – low priority processes may never execute n Solution: Aging – as time progresses increase the priority of the process AE 4 B 33 OSS Lecture 5 / Page 14 Silberschatz, Galvin and Gagne © 2005

Round Robin (RR) n Each process gets a small unit of CPU time (time

Round Robin (RR) n Each process gets a small unit of CPU time (time quantum), usually 10 -100 milliseconds. After this time has elapsed, the process is preempted and added to the end of the ready queue. n 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. n Performance q large FCFS l q small q must be large with respect to context switch, otherwise overhead is too high l AE 4 B 33 OSS Lecture 5 / Page 15 Silberschatz, Galvin and Gagne © 2005

Example of RR with Time Quantum = 20 Process P 1 53 P 2

Example of RR with Time Quantum = 20 Process P 1 53 P 2 17 P 3 68 P 4 24 Burst Time n The Gantt chart is: P 1 0 P 2 20 37 P 3 P 4 57 P 1 77 P 3 P 4 P 1 P 3 97 117 121 134 154 162 n Typically, higher average turnaround than SJF, but better response AE 4 B 33 OSS Lecture 5 / Page 16 Silberschatz, Galvin and Gagne © 2005

Time Quantum and Context Switch Time Turnaround Time Varies With the Time Quantum AE

Time Quantum and Context Switch Time Turnaround Time Varies With the Time Quantum AE 4 B 33 OSS Lecture 5 / Page 17 Silberschatz, Galvin and Gagne © 2005

Multilevel Queue n Ready queue is partitioned into separate queues: foreground (interactive) background (batch)

Multilevel Queue n Ready queue is partitioned into separate queues: foreground (interactive) background (batch) n Each queue has its own scheduling algorithm foreground – RR l background – FCFS l n Scheduling must be done between the queues Fixed priority scheduling; (i. e. , serve all from foreground then from background). Danger of starvation. l 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 l 20% to background in FCFS l AE 4 B 33 OSS Lecture 5 / Page 18 Silberschatz, Galvin and Gagne © 2005

Multilevel Queue Scheduling AE 4 B 33 OSS Lecture 5 / Page 19 Silberschatz,

Multilevel Queue Scheduling AE 4 B 33 OSS Lecture 5 / Page 19 Silberschatz, Galvin and Gagne © 2005

Multilevel Feedback Queue n A process can move between the various queues; aging can

Multilevel Feedback Queue n A process can move between the various queues; aging can be treated this way n Multilevel-feedback-queue scheduler defined by the following parameters: l l l AE 4 B 33 OSS 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 Lecture 5 / Page 20 Silberschatz, Galvin and Gagne © 2005

Example of Multilevel Feedback Queue n Three queues: l Q 0 – RR with

Example of Multilevel Feedback Queue n Three queues: l Q 0 – RR with time quantum 8 milliseconds l Q 1 – RR time quantum 16 milliseconds l Q 2 – FCFS n Scheduling l A new job enters queue Q 0. When it gains CPU, job receives 8 milliseconds. If it exhausts 8 milliseconds, job is moved to queue Q 1. l At Q 1 the job receives 16 additional milliseconds. If it still does not complete, it is preempted and moved to queue Q 2. AE 4 B 33 OSS Lecture 5 / Page 21 Silberschatz, Galvin and Gagne © 2005

Multiple-Processor Scheduling n CPU scheduling more complex when multiple CPUs are available l Multiple-Processor

Multiple-Processor Scheduling n CPU scheduling more complex when multiple CPUs are available l Multiple-Processor Scheduling has to decide not only which process to execute but also where (i. e. on which CPU) to execute it n Homogeneous processors within a multiprocessor n Asymmetric multiprocessing – only one processor accesses the system data structures, alleviating the need for data sharing n Symmetric multiprocessing (SMP) – each processor is self-scheduling, all processes in common ready queue, or each has its own private queue of ready processes n Processor affinity – process has affinity for the processor on which it has been recently running Reason: Some data might be still in cache l Soft affinity is usually used – the process can migrate among CPUs l AE 4 B 33 OSS Lecture 5 / Page 22 Silberschatz, Galvin and Gagne © 2005

Thread Scheduling n Local Scheduling – How the threads library decides which thread to

Thread Scheduling n Local Scheduling – How the threads library decides which thread to put onto an available LWP n Global Scheduling – How the kernel decides which kernel thread to run next n Pthreads library has calls to choose different scheduling policies and parameters AE 4 B 33 OSS Lecture 5 / Page 23 Silberschatz, Galvin and Gagne © 2005

Windows XP Priorities Relative priorities within each class Priority classes (assigned to each process)

Windows XP Priorities Relative priorities within each class Priority classes (assigned to each process) Relative priority “normal” is a base priority for each class – starting priority of the thread l When the thread exhausts its quantum, the priority is lowered l When the thread comes from a wait-state, the priority is increased depending on the reason for waiting l 4 A thread released from waiting for keyboard gets more boost than a thread having been waiting for disk I/O AE 4 B 33 OSS Lecture 5 / Page 24 Silberschatz, Galvin and Gagne © 2005

Linux Scheduling n Two algorithms: time-sharing and real-time n Time-sharing Prioritized credit-based – process

Linux Scheduling n Two algorithms: time-sharing and real-time n Time-sharing Prioritized credit-based – process with most credits is scheduled next l Credit subtracted when timer interrupt occurs l When credit = 0, another process chosen l When all processes have credit = 0, recrediting occurs l 4 Based on factors including priority and history n Real-time Soft real-time l POSIX. 1 b compliant – two classes l 4 FCFS and RR 4 Highest priority process always runs first AE 4 B 33 OSS Lecture 5 / Page 25 Silberschatz, Galvin and Gagne © 2005

Real-Time Systems n A real-time system requires that results be not only correct but

Real-Time Systems n A real-time system requires that results be not only correct but in time l produced within a specified deadline period n An embedded system is a computing device that is part of a larger system l automobile, airliner, dishwasher, . . . n A safety-critical system is a real-time system with catastrophic results in case of failure l e. g. , railway traffic control system n A hard real-time system guarantees that real-time tasks be completed within their required deadlines l mainly single-purpose systems n A soft real-time system provides priority of real-time tasks over non real-time tasks l AE 4 B 33 OSS a “standard” computing system with a real-time part that takes precedence Lecture 5 / Page 26 Silberschatz, Galvin and Gagne © 2005

Real-Time CPU Scheduling n Periodic processes require the CPU at specified intervals (periods) n

Real-Time CPU Scheduling n Periodic processes require the CPU at specified intervals (periods) n p is the duration of the period n d is the deadline by when the process must be serviced (must finish within d) – often equal to p n t is the processing time AE 4 B 33 OSS Lecture 5 / Page 27 Silberschatz, Galvin and Gagne © 2005

Scheduling of two tasks (N = number of processes) Can be scheduled if r

Scheduling of two tasks (N = number of processes) Can be scheduled if r – CPU utilization Process P 1: service time = 20, period = 50, deadline = 50 Process P 2: service time = 35, period = 100, deadline = 100 When P 2 has a higher priority than P 1, a failure occurs: AE 4 B 33 OSS Lecture 5 / Page 28 Silberschatz, Galvin and Gagne © 2005

Rate Monotonic Scheduling (RMS) n A process priority is assigned based on the inverse

Rate Monotonic Scheduling (RMS) n A process priority is assigned based on the inverse of its period n Shorter periods = higher priority; n Longer periods = lower priority n P 1 is assigned a higher priority than P 2. Process P 1: service time = 20, period = 50, deadline = 50 Process P 2: service time = 35, period = 100, deadline = 100 works well AE 4 B 33 OSS Lecture 5 / Page 29 Silberschatz, Galvin and Gagne © 2005

Missed Deadlines with RMS failure Process P 1: service time = 25, period =

Missed Deadlines with RMS failure Process P 1: service time = 25, period = 50, deadline = 50 Process P 2: service time = 35, period = 80, deadline = 80 N 2 3 4 5 10 20 RMS is guaranteed to work if N = number of processes sufficient condition AE 4 B 33 OSS Lecture 5 / Page 30 0, 828427 0, 779763 0, 756828 0, 743491 0, 717734 0, 705298 Silberschatz, Galvin and Gagne © 2005

Earliest Deadline First (EDF) Scheduling n Priorities are assigned according to deadlines: the earlier

Earliest Deadline First (EDF) Scheduling n Priorities are assigned according to deadlines: the earlier the deadline, the higher the priority; the later the deadline, the lower the priority. Process P 1: service time = 25, period = 50, deadline = 50 Process P 2: service time = 35, period = 80, deadline = 80 Works well even for the case when RMS failed PREEMPTION may occur AE 4 B 33 OSS Lecture 5 / Page 31 Silberschatz, Galvin and Gagne © 2005

RMS and EDF Comparison n RMS: Deeply elaborated algorithm l Deadline guaranteed if the

RMS and EDF Comparison n RMS: Deeply elaborated algorithm l Deadline guaranteed if the condition is satisfied (sufficient condition) l Used in many RT OS l n EDF: Periodic processes deadlines kept even at 100% CPU load l Consequences of the overload are unknown and unpredictable l When the deadlines and periods are not equal, the behaviour is unknown l AE 4 B 33 OSS Lecture 5 / Page 32 Silberschatz, Galvin and Gagne © 2005

End of Lecture 5 Questions?

End of Lecture 5 Questions?