Models for addressing cache coherence problems in multilevel caches: A review paper – I. I. Arikpo and A. T. Okoro
ABSTRACT
The computational tasks of the 21st century require systems with high processing speed. This can be achieved by having, among other system specifications, a high capacity microprocessor and large cache size in a single processor system. The cache prevents round-trips to RAM for every memory access. Nevertheless, dependence on a single cache for both data and instructions limits the performance of the system regardless of the capacity of the processor and cache size. Multilevel cache is a technological advancement that has made it possible for a multi-core processor system to have more than one cache. The introduction of multilevel cache technology no doubt speeds up system operations and enhances performance, but is also not without limitations. Cache coherence issues are common problems inherent in this design. Different protocols have been designed to ameliorate the issues surrounding cache coherence. The aim of this paper is to discuss the efficiency of these protocols in addressing cache coherence problems in multilevel cache technologies.