A Comparison of Three Genetic Algorithms for Locking-Cache Contents Selection in Real-Time Systems

Abstract

Locking caches, providing full determinism and good performance, are a very interesting solution to replacing conventional caches in real-time systems. In such systems, temporal correctness must be guaranteed. The use of predictable components, like locking caches, helps the system designer to determine if all the tasks will meet its deadlines. However, when locking caches are used in a static manner, the system performance depends on the instructions loaded and locked in cache. The selection of these instructions may be accomplished through a genetic algorithm. This paper shows the impact of the fitness function in the final performance provided by the real-time system. Three fitness functions have been evaluated, showing differences in the utilisation and performance obtained.

Impact of L1 entire locking and L2 way locking on the performance, power consumption, and predictability of multicore real-time systems

“…In [17], a methodology to select a set of instructions to be preloaded in the cache using a genetic algorithm is proposed. In [18], various algorithms to select a set of instructions to be locked in cache are compared. These algorithms show better performance and simultaneously estimate a tight upper bound of the response time of tasks.…”

“…These algorithms show better performance and simultaneously estimate a tight upper bound of the response time of tasks. Techniques discussed in [14][15][16][17][18] are used mainly to evaluate predictability in a single-core system. These techniques are not capable of power estimation -a crucial design factor for embedded systems.…”

“…Therefore, it becomes a great challenge to support real-time applications on multicore systems. Studies show that for single-core systems, cache locking improves predictability [8][9][10][11][12][13][14][15][16][17][18]. Cache locking is the ability to prevent some or all of the instruction cache (I1) or data cache (D1) from being overwritten.…”

An algorithm for deciding minimal cache sizes in real-time systems

“…This is the sequential version of the genetic algorithm presented in [15]. Eight different versions have been evaluated.…”

“…Although it is possible to use other metrics such as task slack or response times, in [16] it is shown that using the system utilisation in the fitness function offers similar results than using these aforementioned metrics.…”