In a time-critical system, it must be ensured that all the tasks will meet their deadlines, which are usually dictated by physical constraints.
The execution time of a task depends on several factors including its input data (often read from sensors), the state of the hardware (e.g. the cache contents) when the task starts executing, or conflicts with other tasks when accessing shared resources.
To schedule tasks so that timing constraints are guaranteed to be fulfilled, it is needed to know the worst-case execution times (WCET) of each task. This course will give an insight into how WCETs can be determined.
More precisely, we will discuss the following topics: (i) challenges faced by WCET estimation; (ii) approaches to WCET estimation, from measurement-based methods to static analysis techniques; (iii) models of hardware components, such as the processor pipeline, cache memories or dynamic branch predictors for static WCET analysis; (iv) timing analysis of systems running on multicore platforms.
Christine Rochange is a full professor at the University of Toulouse, France. She is leading the TRACES research group at the IRIT lab (Institut de Recherche en Informatique de Toulouse).
Her main research interests are on static worst-case execution time analysis, in particular on modelling the behaviour of hardware components. Her group is developing OTAWA, an open-source toolbox dedicated to WCET estimation, that has been successfully used in several industrial and academic projects. She is a member of HiPEAC.