Benchmarking is the act of running a computer program or operation in order to compare operations in a way that produces quantitative results, normally by running a set of tests and trials against them.

Benchmarks can be organized in the following two big categories:

Macrobenchmarks exist as a means to compare different platforms in specific areas such as processor speed, number of floating-point operations per unit of time, graphics and 3D performance, and so on. They are normally used against hardware components, but can also be used to test software-specific areas, such as compiler optimization or algorithms.

As opposed to these traditional macrobenchmarks, a microbenchmark attempts to measure the performance of a very small piece of code, often a single method. The results obtained are used to choose between competing implementations that provide the same functionality, when deciding the optimization path.

The risk here is to microbenchmark...