core/renderbackend: improve handling of multiple render time queries

With multiple OpenGL render time queries, you can't just add the render times together, because the GPUs execute commands asynchronously and you may get a far too big result this way. Instead, this converts all timestamps to steady_clock and correctly calculates the actual total time spent rendering with the first and last timestamp.

This should reduce latency on multi gpu setups (on my desktop, I get up to 3ms lower render times than git master would have you believe), make render time tracking more reliable for !4833 (merged) and help with compositing on multiple drm planes later on as well.

closes #209 (closed)