DSP Architecture Design Essentials

This book addresses the gap between DSP algorithm design and implementation. It uses an approach to DSP architecture design that merges algorithm and system specifications with the intrinsic computational capability of the underlying technology. The material includes detailed treatment of DSP implementation with arithmetic, micro-architecture, and circuit-level techniques mapped in energy-area-performance space that serves as a tool for comparing multiple architectural realizations. The architecture optimization procedure is carried out in the graphical Matlab/Simulink environment because most algorithm developers are familiar with Matlab and it also provides a hardware description format that is convenient for circuit designers. This vertical integration approach will be the key differentiating factor of this book.§Another key difference is that has been accomplished via close interaction and collaboration with leading industrial researchers in this area. So the book brings in commercial perspectives and addresses the learning objectives of working professionals. Many examples are presented using Synplify DSP tools. The book covers the fundamentals of development of various DSP hardware platforms; provides scientific treatment of architecture optimization using integer linear programming; analyzes current radio systems; and also includes a discussion about emerging problems related to parallel data processing in applications such as MIMO communications and neuroscience.