Introduction

Abstract

For many years, computer performance improvement was based on technological innovations that allowed to dramatically increase the chip’s transistor count. Moreover, architectural progress aimed at organizing processors structure have allowed to overcome traditional sequential execution of programs by exploiting instruction level parallelism. Yet, last decade has shown that Moore’s Law is reaching its natural breaking point and maintaining the performance improvement rate by decreasing transistor’s size will no longer be possible. Main manufacturers have thus decided to offer more processor kernels in a single chip, opening the way to the multi core era. Examples of that are: the Intel core i3 (2 cores), i5 (4 cores), and i7 (4 cores) architectures, AMD Zambezi, phenom iii (8 cores), phenom ii (6 cores).