Performance profiling
The next set of performance analysis tools that we are going to learn about is the profiling tools, or profilers. We have already seen a profiler in use: in the last section, we used it to identify the function that was taking the majority of the computation time. This is exactly what profilers are used for, to find "hot" functions and code fragments, that is, the lines of code where the program spends the most time.
There are many different profiling tools available, both commercial and open source. In this section, we are going to examine two profilers that are popular on Linux systems. The goal is not to make you an expert on a particular tool but to give you an idea of what to expect from the profiler you choose to use and how to interpret its results.
First, let us point out that there are several different types of profilers:
- Some profilers execute the code under an interpreter or a virtual machine and observe where it spends the...
