Estimation of probabilistic worst case execution time while accounting OS costs

Walid Talaboulma (1) , Cristian Maxim (1, 2) , Adriana Gogonel (1) , Yves Sorel (1) , Liliana Cucu-Grosjean (1)

1 AOSTE - Models and methods of analysis and optimization for systems with real-time and embedding constraints
2 Airbus Operation S.A.S.

The arrival of modern and more complex processors (e.g., use of caches, multi-and many-core processors) increases the timing variability of tasks, i.e., the worst case execution time (WCET) is becoming significantly larger, while the probability of appearance of a worst case execution time is extremely low. Approaches taking into account this probability have been the topic of last years research by defining the probabilistic worst case execution time of a task as a probabilistic bound on all execution time scenarios. Existing measurement-based approaches consider the execution of the tasks in isolation and our contribution provides hints for accounting OS costs.

Domaines: Systèmes embarqués

Walid Talaboulma, Cristian Maxim, Adriana Gogonel, Yves Sorel, Liliana Cucu-Grosjean. Estimation of probabilistic worst case execution time while accounting OS costs. 21st IEEE Real-Time Embedded Technology and Applications Symposium, Apr 2015, Seattle, United States. ⟨hal-01298737⟩