Benchmarks

[pfeatherstone]

Maybe it would be a good idea to use something like https://github.com/google/benchmark to benchmark dlib functionality on every release, or even every commit.
We could add a Github Actions workflow which would run the benchmarks and store the results somewhere, maybe even updating some graphs or CSV tables.

[davisking]

I don't really want to add a dependency. And it's easy to time stuff with just the standard library.

I doubt github actions gives reliable timing results. Whenever I see stuff run on these kinds of cloud services the timings are all over the place. I would just have something in dlib/tools if you want that runs a benchmark and you can then run it locally on a machine you know isn't running anything else and is a sensible reference.

[pfeatherstone]

#2876