Examining Cache Handling of the FIVE Method on Multicore Systems

Abstract

Multicore systems need a new point of view of software engineering when using parallel computing methods. In this case the processor core cache handling was examined in multicore systems for embedded system application. The upper limit of number of rule bases for the FIVE (Fuzzy Interpolation in Vague Environment) method per processor core was determined according to the cache size. The upper limit gives the rulebase size, which can be computed without reading of the system external memory. This can reduce computing time and can help to reach a higher speed-up on a multi core system. The speed-up was calculated with the Amdahl formula.