Added an internal profiler to the project today. I ended up using the Shiny Profiler (found at SourceForge) because it is lightweight, powerful and really simple to set up. It took me 30 minutes to get it to work, compiling included, and then another 30 minutes to get the results I wanted.

It uses some nice preprocessor magic to settle itself in your code so that you only need to call PROFILE_FUNC() at the beginning of each function you want to time. Beyond that you need to call an update function at the end of every frame, or loop or whatever you're measuring over. Finally you call a third function to get a formatted tree like in the pretty picture. There are more advanced options available but those three calls are all you need to get going.

I can really recommend this if you want a simple yet flexible, open-source profiler.

The download says there are no examples but it does include two very simple ones.