what doesn't make sense is, a game that needed optimizations like tomb raider or total war, still had usage across all cores, but gpu usage was low even with multiple cores being used (and not being maxed out)
this seems entirely different in the sense of only a single core is being used, surely the scheduler should be assigning the threads regsrdless?
here's another conundrum for you guys.
3d mark firestrike on my 5960x
http://www.3dmark.com/fs/12008594
here it is again with my 7700k (titan Xp instead of 1080ti)
http://www.3dmark.com/fs/12829241
and finally on my 1800x system
http://www.3dmark.com/fs/12728829
notice how the physics score scales as expected, but gpu gets held back on the ryzen system, worst is the combined score, it takes a massive hit on the 1800x.