I recently did some tests by swapping round components in my two gaming systems. I currently have an FX8350 with a r9 290 and an A8-5600k with a hd 7870. Now my more powerful rig is plugged into a monitor where as my HTPC is plugged into a huge projector screen. I wanted to see whether it would be worth putting my r9 290 into my HTPC. So here are my results:
Unigine heaven: A8-5600k w/7870 @27fps, a8-5600 k w/290 @53fps, 8350 w/290 @54.1fps
Batman Arkham origins: A8 w/ 7870 @65fps, a8 w/290 @112fps, 8350 w/290 @103fps
F1 2012: a8 w/ 7870 @75fps, a8 w/290 @83fps, 8350 w/290 @91fps
Firestrike: a8 w/7870 @4709, a8 w/290 @7007, 8350 w/290 @ 8531
As you can see the bottleneck is small at the Most and in Batman Arkham Origins, the a8-5600k actually somehow beat the 8350 in the built in Benchmark. Both systems used exactly the same RAM (the same actual sticks) Hope this helps and people start to realise that bottlenecks are not as bad as people think. My results actually reflect those that were achieved by JayztwoCents on YouTube.