The thesis explores how different cache replacement policies affect the accuracy and reliability of Worst-Case Execution Time (WCET) analyses. It delves into the relationship between these policies and their impact on performance metrics, aiming to enhance understanding of how cache management strategies can optimize timing predictions in computational systems.
Embedded systems as they occur in application domains such as automotive, aeronautics, and industrial automation often have to satisfy hard real-time constraints. Safe and precise bounds on the worst-case execution time (WCET) of each task have to be derived. This thesis studies the influence of cache replacement policies on the precision and soundness of WCET analyses.