Carrizo/Bristol Ridge have significantly faster L2, and in a majority of non-gaming benchmarks, the XV-based CPUs whip the SR-based CPUs soundly. In a fair number of games, the opposite is true.
Well most of the gaming comparisons weren't "normalized" to the same mhz.
3.5 Ghz XV base clock vs 3.7/3.9/4.0 Ghz....hardly seemed professional.
I do remember that the total average mhz for mhz difference was like 2-3% in gaming for Kaveri/Godavari if I recall correctly.
Though I'm not sure this will still hold up...aside from that XV on AM4 will sit at least on 3.8 Ghz base for their "top dog" for now...might even be higher, not like we already know the final specs of the DIY Chips. (960(K),970(K) and so on)
In addition to that we get different ram and updated Chipset...these should all play in Bristol Ridges' favor and at least remove those pesky 3% average hz for hz gaming "loss".
Not sure how much
This wccftech article is to be trusted...but at least there it looks like that even the shitty OEM version seems to outperform Kaveri on all ends. But then again...the test there is also hella awkward. 1600mhz DDR3 vs 2133mhz SINGLE CHANNEL DDR4...just a big plain "wat?" xD