fakultät für informatik informatik 12 technische universität dortmund Lab 3: Scheduling - Session 10 - Peter Marwedel Heiko Falk TU Dortmund Informatik 12 Germany

- 2 - technische universität dortmund fakultät für informatik  p. marwedel, informatik 12, 2008 TU Dortmund Rate monotonic scheduling  1. Exercise: Assume a system of two periodic tasks. Task 1 has a period of 5 units and an execution time of 3. Task 2 has a period of 8 and an execution time of 3. Let the deadline be equal to its period. a)Using rate monotonic scheduling, can any of the two processes miss its deadline? Construct a formal argumentation answering this question, based on the formula about the utilization of a processor when using rate monotonic scheduling. b)Which schedule is generated by rate monotonic scheduling for the above two tasks? Draw a diagram where the x-axis reflects the time and the y-axis lists all tasks. Put vertical lines in the diagram at those points of time where tasks become available. Draw rectangles for those times where tasks are really executed. Mark all those points of time where deadlines are missed, if this ever happens. When scheduling, assume that the execution of a task that has just missed its deadline is not caught up. c)Verify your drawn schedule by using the software leviRTS.

- 3 - technische universität dortmund fakultät für informatik  p. marwedel, informatik 12, 2008 TU Dortmund EDF  2. Exercise: Consider the following set of tasks where A n denotes the arrival time, d n denotes the deadline and c n denotes the execution time: a)Create a schedule of the above tasks using least laxity. Draw a diagram like in exercise 1 showing when which task becomes active. Is there a task missing its deadline? b)Create a schedule of the above tasks using earliest deadline first. Draw a diagram like in exercise 1 showing when which task becomes active. Is there a task missing its deadline? c)Verify your EDF schedule by making leviRTS produce the desired schedule. AnAn dndn cncn T1T T2T T3T T4T

- 4 - technische universität dortmund fakultät für informatik  p. marwedel, informatik 12, 2008 TU Dortmund Latest Deadline First  3. Exercise: Assume a system of five interdependent tasks T 1 to T 5. The dependencies between the tasks are described by the task graph below (Note: an edge T i  T j means that task T j can only be executed if T i has finished execution). c n denotes the execution time of a task and d n denotes the deadline interval of a task. T 1 with c 1 = 2 and d 1 = 15, T 2 with c 2 = 5 and d 2 = 20, T 3 with c 3 = 4 and d 3 = 12, T 4 with c 4 = 3 and d 3 = 15, T 5 with c 5 = 3 and d 5 = 20 Schedule this task set using latest deadline first. T2T2 T3T3 T1T1 T4T4 T5T5

- 5 - technische universität dortmund fakultät für informatik  p. marwedel, informatik 12, 2008 TU Dortmund Resource Access Protocols  4. Exercise: A task set consisting of tasks T 1, T 2, T 3 and T 4 should be executed on a processor with the following priorities: P 1 = 4 (lowest), P 2 = 3, P 3 = 2, P 4 = 1 (highest). The table shows the values for the arrival times, execution times and the resource accesses for each task. a)Generate an empty task scenario in the software leviRTS and select the algorithm Resource Access Protocol (priority based, preemptive). Create the tasks T 1 to T 4 with the above properties. b)Start the visualization of this scenario. Which problems occur during scheduling? What can be done to solve them? c)Draw (manually) a schedule of the above tasks under the assumption that priority inheritance is used. Verify your drawn schedule using leviRTS. TaskArrival Time Execution Time PrinterCOM1  t P  t V  t P  t V T1T T2T T3T T4T