Introduction

In an ideal situation, we can select an embedded processor for our application that provides maximum performance for minimum extra development effort. In this utopian environment, we could code everything in a high-level language like C, we wouldn't need an intimate knowledge of our chosen device, it wouldn't matter where we placed our data and code, we wouldn't need to devise any data movement subsystem, the performance of external devices wouldn't matter In short, everything would just work.

Alas, this is only the stuff of dreams and marketing presentations. The reality is, as embedded processors evolve in performance and flexibility, their complexity also increases. Depending on the time-to-market for your application, you will have to walk a fine line to reach your performance targets. The key is to find the right balance between getting the application to work and achieving optimum performance. Knowing when the performance is "good enough" rather than optimal can mean getting your product out on time versus missing a market window.

In this chapter, we want to explain some important aspects of processor architectures that can make a real difference in designing a successful multimedia system. Once you understand the basic mechanics of how the various architectural sections behave, you will be able to gauge where to focus your efforts, rather than embark on the noble yet unwieldy goal of becoming an expert on all aspects of your chosen processor.

We discussed the basic components of the Blackfin architecture in Chapter 1. Here,...