1. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.1 Operating System Concepts Operating Systems Lecture 17 Scheduling III

2. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.2 Operating System Concepts Recall Scheduling Criteria CPU utilization –fraction of time the CPU is busy. CPU efficiency – fraction of time the CPU is executing user code. Throughput – # of processes completed per unit time Average Turnaround time – average delay between job submission and job completion. Normalized turnaround time – Ratio of turnaround time to service time per process. Indicates the relative delay experienced by a 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.

3. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.3 Operating System Concepts ProcessArrival TimeBurst Time P 1 0.07 P 2 2.04 P 3 4.01 P 4 5.04 SJF (non-preemptive) Selection criterion: min(s) Average waiting time = (0 + 3 + 6 + 7)/4 = 4 Avg service time = (7 + 4 + 1 + 4)/4 = 4 Throughput = 4/16 = 0.25 Avg turnaround = (7 + 10 + 4 + 11)/4 = 32/4 = 8  Check consistency. Wait = (Turnaround - Service - Dispatch) Example of Non-Preemptive SJF P1P1 P3P3 P2P2 73160 P4P4 812

4. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.4 Operating System Concepts Example of Preemptive SJF (SRTF-- Shortest remaining time first) ProcessArrival TimeBurst Time P 1 0.07 P 2 2.04 P 3 4.01 P 4 5.04 SJF (preemptive) Selection criterion: min(s - e) Average waiting time = (9 + 1 + 0 + 2)/4 = 3 What statistics are different from non-preemptive SJF? What are the values of these stats? P1P1 P3P3 P2P2 42 11 0 P4P4 57 P2P2 P1P1 16

5. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.5 Operating System Concepts Determining Length of Next CPU Burst Can be done by using the length of previous CPU bursts, using exponential averaging. If we expand the formula, we get:  n+1 =  t n +(1 -  )  t n -1 + … +(1 -  ) j  t n -j + … +(1 -  ) n+1  0

6. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.6 Operating System Concepts Priority Scheduling A priority number (integer) is associated with each process. The CPU is allocated to the process with the highest priority  Some systems have a high number represent high priority.  Other systems have a low number represent high priority.  Text uses a low number to represent high priority. Priority scheduling may be preemptive or nonpreemptive.

7. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.7 Operating System Concepts Assigning Priorities SJF is a priority scheduling where priority is the predicted next CPU burst time. Other bases for assigning priority:  Memory requirements  Number of open files  Avg I/O burst / Avg CPU burst  External requirements (amount of money paid, political factors, etc). Problem: Starvation -- low priority processes may never execute. Solution: Aging -- as time progresses increase the priority of the process.

8. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.8 Operating System Concepts Round Robin Scheduling Each process gets a small unit of CPU time (time quantum).  A time quantum is usually 10-1000 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.

9. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.9 Operating System Concepts Example of RR, time quantum = 20 ProcessBurst Time P 1 53 P 2 17 P 3 68 P 4 24 The Gantt chart is: Compute: Avg service time, Throughput, avg turnaround, avg wait: Typically, higher average turnaround than SJF, but better response. Suppose P 1 arrives at 0, P 2 at 19, P 3 at 23 and P 4 at 25. What changes? What are the new values? P1P1 P2P2 P3P3 P4P4 P1P1 P3P3 P4P4 P1P1 P3P3 P3P3 02037577797117121134154162

10. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.10 Operating System Concepts RR Performance Performance varies with the size of the time slice, but not in a simple way. Short time slice leads to faster interactive response.  Problem: Adds lots of context switches. High Overhead. Longer time slice leads to better system throughput (lower overhead), but response time is worse.  If time slice is too long, RR becomes just like FCFS. Time slice vs process switch time:  If time slice = 20 msec and process switch time = 5 msec, then 5/25 = 20% of CPU time spent on overhead.  If time slice = 500 msec, then only 1% of CPU used for overhead.  The time slice should be large compared to the process switch time.  A typical time slice is 1 sec (4.3 BSD UNIX) RR makes the implicit assumption that all processes are equally important.  Cannot use RR is you want different processes to have different priorities.

11. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.11 Operating System Concepts Multilevel Queues Ready queue is partitioned into separate queues: E.g.: foreground (interactive) background (batch) Each queue has its own scheduling algorithm, e.g.: foreground – RR background – FCFS Scheduling must be done between the queues. Possible methods:  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 Works well for time sharing systems.

12. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.12 Operating System Concepts Multilevel Queue scheduling

13. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.13 Operating System Concepts 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 (where to put new processes)

14. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.14 Operating System Concepts Typical behavior for multilevel feedback queues New processes go to the highest priority queue for that job type.  Top priorities reserved for system processes. If process uses full time slice, it moves down a priority. If process blocks before using full time slice, it remains at the same priority. Higher priority queues have smaller time slices than lower priority queues.

15. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.15 Operating System Concepts Example of a Multilevel Feedback Queue Three queues:  Q 0 – time quantum 8 milliseconds  Q 1 – 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.

16. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.16 Operating System Concepts Multilevel Feedback Queues

17. Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, 2005 6.17 Operating System Concepts Notes on Multilevel Feedback Queues Short processes are favored. Good for interactive processes with short CPU bursts. Is starvation Possible?  Yes--Long processes may wait forever.  To avoid--increase the priority of a process if it has been waiting for some period of time in some queue.