The way this sort of thing is typically done is to work with the performance measurement people for your system and identify which parts of the code use the most resources (if there is an interest in this level of performance measurement, they will have tools to use to measure). Once the "heavy" user sections of code are identified, they can be reviewed to see how they process and how they might possibly be improved.
Most performance improvements are made by changing the way i/o is done or the way database activity is done.
Keep in mind that it is more important that the code always executes correctly and that the code is maintainable.
Hope this helps,