Performance measurement is often subject to many debates, but we should approach the ways of solving performance problems as straightforwardly as possible while maintaining objective processes. The results must be as objective as they can. To correctly define that a performance optimization is needed or not, we must be able to measure the running code objectively. To ensure the objectiveness of the performance measurement, the result must be visible as quantitative (in numbers) and qualitative by analyzing how the code behaves when it runs, how fast it runs, and how big the code is in memory.
As a rule of thumb, it is easier to analyze quantitatively as data can be seen and compared more directly than when analyzed qualitatively. Understanding how to measure and how to interpret the measurement result can be used as a foundation for deducing the cause of any performance bottlenecks and can be further used in combination with qualitative analytics, such as...