The lecture notes (in the form of PowerPoint "handouts") will be available on the course home page at least 48 hours before the exercise session in question.
In this exercise, we give an overview of the laboratory assignment, in terms of the main objectives of the assignment and also in terms of the hardware platform and software development tools being used. We also present some special considerations when using the TinyTimber kernel.
In this exercise, we demonstrate how to add the call-back functionality to the mutex methods provided by the TinyTimber kernel. We then show how to implement semaphores in C using the TinyTimber kernel and the call-back functionality add-on.
In this exercise, we demonstrate how to implement periodic tasks in TinyTimber. In addition, we will show how to impose deadlines to periodic tasks and how to stop a periodic task after it has executed for a certain number of time units.
In this exercise, we will learn how estimate the worst-case execution time (WCET) for a given piece of code using Shaw's method.
In this exercise, we will solve problems related to pseudo-parallel execution on a uniprocessor with the Rate-Monotonic (RM) and Earliest-Deadline First (EDF) scheduling policies, including the use of processor utilization analysis. In addition, we will learn how to generate a time table for a system that uses a cyclic executive.
In this exercise, we will solve uniprocessor scheduling problems related to the Deadline-Monotonic (DM) and Earliest-Deadline First (EDF) scheduling policies. The solutions to the problems are based on response-time analysis for DM and processor-demand analysis for EDF.
In this exercise, we will solve multiprocessor scheduling problems. The solutions to the problems are based on processor utilization analysis for the RMFF and RM-US scheduling policies. We will learn that successfully assigning the tasks to the processors using the RMFF scheduling policy depends on the order of the tasks considered during the assignment process. We will also demonstrate how to compute a pessimistic upper bound on the response time of tasks scheduled using the RM-US policy.
Printable version