General Principles

One of the primary goals of performance tuning is to find the areas of code that are used the most in an application and reduce the amount of the time it takes to process those areas. Although other areas can be considered, such as reducing memory or other resource requirements, we will concern ourselves in this section with speed as the primary measure of efficiency.

Many factors affect performance. Some of them are not under the application programmer's control, such as machine architecture. Others, such as how I/O and memory are used, are directly observable (and tunable) using the concepts described in this chapter. A few are basic to the construction of the applications and require serious consideration both before development begins and after performance analysis. These include the algorithms used as the basis of programming or the size and granularity of the tasks being performed.

To target those programs most in need of optimization, profiling tools are available. For our next discussion, we'll take a look at profiling.