It makes sense for Nintendo as they haven't had a new system architecture for over a decade. They are ripe for a new system now, quite possibly ARM based and compatible with their handhelds for product synergy if they decide best power isn't of value to them (as it has proven with Wii and their handhelds).
I remember in one of the ask Iwata talks they preferred systems which didn't have as many 'gotchas' in terms of running a variety of code where one small change absolutely tanks framerate. Now that I think about it, it supports fewer cores/threads with higher performance per core. Does that bring X86 front and centre to the playing field? I don't know enough about ARM to know how it compares under this scenario.
P.S. Where do you send the mod bribes to again?
CPU wise, probably barely faster or even slower. There was a talk some page ago about how Bobcat doesn't even competitive with XCPU in single-thread performance.. Krishna should have an enhanced version of Bobcat, but we don't expect miracles.
You have to remember that CISC/RISC flops aren't directly comparable because of how the architectures handle instructions. You'd also have to consider which process node and which foundry they would use to produce the CPU.
It wouldn't solve one of the biggest problem of Nintendo Wii: the lack of good third party support.
You'd expect that the current generation consoles would linger well after the next generation enters the picture. I don't think they'd have any problems so long as they have more memory than the current generation consoles as well as enough performance to be close enough to the next generation consoles.
We could probably use the info we have on the 360 design & manufacturing timeline to get an idea of how rapid things can be.
They'd gain a fair bit of time simply through using an X86 architecture on that front, and if they use ARM they could piggyback off early NGP development. Either way both architectures shouldn't need too much lead time with developers with respect to development kits, right?
So sure, two years would be an insanely huge amount of time if the design is done. If it's a cutting edge process, there will be problems with getting good yields - it's not just whether the chip works or not, but also getting chips at the desired clockspeeds (or rather, chips that behave within operational power and thermal designs). It'll take some time to experiment/optimize the chemical process as the console companies can't rely on binning as Intel and AMD do. Still I wouldn't expect that to take too long to fix with that sort of timeline (much less even 6months to a year) as they can focus on fixing these sorts of production issues instead of trying to push out chips as fast as possible.
On a more mature process, the problems should be pretty insignificant unless we're talking monolithic 400+ mm^2 chips. Then it's just a matter of ramping up production.
Aren't yields also a function of voltage as well as clockspeed? It's not just a question of how many chips simply work but also how many work at the set clockspeed and thermal parameters? So if worse comes to the worse they could go with say a vapour chamber cooler initially and then spec it down to a regular heatsink/fan combo when yields improve and required voltages drop?
One burning question I have is whether they could possibly get a console out in 2011 as those rumours suggest or early 2012 even on 28nm since we now have two contract fab companies slated to offer 28nm production this year? Thats the real question since it would have an extreme bearing on what they could pack in at their desired thermal limit. Although in saying that 32nm SOI is also a possible candidate although I don't know how much capacity GF can spare to anyone who isn't called AMD!
Anyway, which node do you think?