Optimizing Compilers for Embedded Appli ations

In reasingly, general-purpose pro essors are used for embedded appli ations rather than ustomised hardware. As pro essor ost drops, it be omes more attra tive to use one pro essor for several appli ations rather than designing spe i hardware. Multimedia based appli ations are typi al of the growing uses of embedded systems, requiring ost-e e tive implementation and high performan e. Very Long Instru tion Word (VLIW) pro essors are an attra tive solution for su h appli ations as they provide potentially high performan e, due to multiple parallel fun tional units, and are relatively heap to manufa ture due to the simple pro essor ar hite ture. However, sophisti ated optimizing ompiler te hnology is ne essary to exploit the ne-grain parallelism as assembly programming of omplex appli ations is not feasible.