Lets hope its more impressive than Ivy bridge. Ivy was more of a step backwards than forwards. 22nm did nothing for the desktop market
I doubt it'll get better. If I was a betting man, I'd say it would get worse and worse as it goes on.
They can't do both, focus on impressively advancing on the desktop, while holding off ARM on the low power space.
I'm expecting:
-10-15% better performance
-Similar air overclock as Ivy Bridge
-Better top overclocks using exotic cooling(but this is due to other factors like increasing the max multiplier and giving different base clock frequencies)
It's especially true because 22nm breaks the tradition of increasing drive currents to be high as possible. It doesn't, instead it focuses on lowering voltages and perf/watt(IOW, better performance at lower voltages).
In my opinion, that was the biggest sign that Intel was being serious about competing in the lower power market: By changing the nature of their process technology, their most important asset. There are bound to be sacrifices, and I wouldn't be surprised it was with the Desktop chips.