Lieu : ISAE-SUPAERO, Salle 11.126 (Bâtiment 11 - DFM) Et en ligne via zoom : https://zoom.us/j/98180404932?pwd=bWhDeHdLTnhCN2RLdEs2QVJxUWtMQT09
Memory interference caused by resource sharing on multicore platforms lead to execution
overheads, which is a security issue for safety-critical applications. To ensure that a given task
is not exceeding its execution deadline, predictability and interference impact must be
optimized. One way to achieve this objective is through task/memory mapping, which consists
in associating a task to a core and the task memory to a DDR memory bank. These associations
can be optimally obtained using metaheuristic multi-objective optimization algorithms. A part
from the DDR memory interference optimization, other criteria can be considered at the same
time, e.g., tasks execution parallelism, energy consumption.