14nm++ per Intel claims should bring the same increase in performance as 14nm+ from the original 14nm ( around 12% better performance in ideal conditions ).
Ideal, so it'll turn out to be far less. I wonder if even 100MHz increase at the same core is in order? Of course I'm talking about overclocking frequency. Overclocking headroom will decrease because base clocks might be upped more.
Ideal as in:
-What metric are they talking about? At 0.7v, which is practically idle?
-Probably when the clock isn't limited by being near the limits of frequency(~5GHz)?
-It doesn't have some quirks in the circuit that limits clocks
Its a real bad time to be in the high-end CPU design. Most of the consumer base heavily relies on clocks and perf/clocks, they don't scale with cores, yet that's the only thing you can do because you can't increase clocks anymore. Even if you could increase it massively most people don't need the extra performance.