ISA Extension to General Purpose Processors for Fuzzy and Genetic Algorithm Applications

The number of intelligent systems which use soft-computing methods in real-world applications is growing. General purpose processors (GPPs) are one selection of designers to implement intelligent systems. However, to accelerate the performance of them, an application-specific instruction set extension is added to the basic processor in some designs. In this paper, fuzzy logic and genetic algorithm (GA) applications are selected as two soft-computing techniques in the domain of intelligent systems which their supportive instructions are added to the architecture of LEON2 general purpose processor. The achieved results show great improvements in some intelligent applications' execution times. This ISA extension has added to the GPP's core an area with less than 6% overhead for fuzzy instructions and about 14% for GA instructions.