Professional Embedded ARM Development

Chapter 10

Writing Optimized C

WHAT’S IN THIS CHAPTER?

• Knowing when to optimize
• Knowing what to optimize
• C optimization techniques
• Assembly optimization techniques
• Hardware optimization

Optimization is the final part of any project, and it is vitally important to understand. Some developers write optimized code from the start, and on most projects, that is a bad idea. The reason is simple: Optimized code is often difficult to read or to understand, and during the development phase, changes have to be made, often from different people, and in some cases, different departments. It is often best to start with readable, maintainable code before starting the optimization process. In addition, spending an extra two hours on optimizing a section of code might not be worth it; maybe this function will be called only once, or maybe the entire file will be replaced when changes occur in the project.

When the project is finished, when all the functionality has been added and all the bugs have been corrected, now it’s time to start optimizing. The problem is, where do you start?

RULES FOR OPTIMIZED CODE

The term rules are used, but the truth is there are no set rules. You do not need to follow every rule here, indeed, for some applications; some of these rules might be impractical. You are the only judge as to what should be used and what shouldn’t. The previous code used 32-bit integers, but in some cases, you would prefer (or need) to have two 16-bit numbers because space could be critical, ...