Such a system is not a slouch and that makes me wonder about what is going on internally.
It's mostly the emulation of graphics.library and Picasso96. All of those components use lots of small functions, and when these are called repeadetly, the performance breaks in because the overhead of the actual emulation switch adds up.
Currently, we're converting the graphics system to PPC. This unfortunately takes time, as it's the most hardware dependant part of the whole system, and we want to keep the AGA compatibility on the classic systems
Let's just say that the video shows there is really a lot of room for improvement
Of course there is. If it wasn't, it would alread be out :-D