PACO: fast average-performance estimation for time-randomized caches

Probabilistic timing analysis is a powerful approach to derive worst-case execution time (WCET) estimates, as needed in safety-critical systems, in the presence of high-performance hardware features (e.g., caches). To that end, the timing behavior of certain hardware resources, such as caches, is randomized. Time-randomized (TR) caches allow deriving hit/miss probabilities for each access and probabilistic WCET estimates for the overall program. However, the analysis of the average performance of TR caches, which is needed for lowly-critical high-performance tasks in mixed-criticality environments, has been neglected. So far, average performance of a TR cache can only be analyzed through simulation, whose accuracy strongly depends on carrying a large number of simulations. In this paper we address this challenge by proposing PACO, an accurate analytical approach to estimate cache hit/miss probabilities of full applications, parts of them and individual cache accesses at low cost for a wide variety of TR cache hierarchies and setups.