ARMs 20nm Doesn't Get the Nod, Even When It Does Arrive (Unknown)
The challenge for both TSMC and GlobalFoundries is going to be how to match the performance of Intel’s 22nm technology with their own 28nm products. 20nm looks like it won’t be able to do so, which is why both companies are emphasizing their plans to move to 16nm/14nm ahead of schedule. There’s some variation on which node comes next; both GlobalFoundries and Intel are talking up 14nm; TSMC is implying a quick jump to 16nm.
Will it work? Unknown. TSMC and GlobalFoundries both have excellent engineers, but FinFET is a difficult technology to deploy. Ramping it up more quickly than expected while simultaneously bringing up a new process may be more difficult than either company anticipates. Given the advantages Intel claims for the technology, it might’ve made more sense to ramp FinFET on an established node. One of the most significant demonstrations of what Intel thinks it’s getting out of 22nm FinFET is the company’s decision to revise Atom for an out-of-order architecture.
22nm Atom should close the gap with existing ARM CPUs and give Intel a substantial advantage. Overall, the situation looks like Intel holds the cards until GF and TSMC manage to revise their roadmaps for the sub-20nm market.
TSMC 20nm has been announced and is on its way. Since 28nm is now in volume production, the obvious transition would be to directly scale that down. Why would they skip that and go straight to 16/14 nm FinFET? It will probably come online I'd say 6 months after Intel's 22nm so Intel will have a short-lived lead on SOC process, the same as it had when it introduced 32nm Medfield.
[LOL. Yeah, it's been "one its way" for two years.]
"Since 28nm is now in volume production, the obvious transition would be to directly scale that down."
[Volume production? You call that volume production? Don't make me laugh. As far as "just scale it down", your ignorance of fabrication never fails to amuse. The problem with non-FinFET 20nm is that it just doesn't get the job done. It costs too much for too little improvement. The economics are bad, bad, bad. And this totally ignores the issue of whether the ARM foundries can produce it and when. It's taken TSMC forever to slowly ramp 28nm. 20nm is going to be harder and more complex.]
"Why would they skip that and go straight to 16/14 nm FinFET? It will probably come online I'd say 6 months after Intel's 22nm so Intel will have a short-lived lead on SOC process, the same as it had when it introduced 32nm Medfield."
[See note above on why they want to move directly to 14/16nm. Your estimate of when 20nm will come online is based on absolutely nothing. No one knows because it's so far out. And volume production is even further out. And volume production that would satisfy Apple even further out than that. It's game over, dude. Intel has won the fabrication war. Only you and Chump Street don't know yet....]
The ARM foundries have struggled mightily with just 28nm production. Now they would have us and customers believe that they can successfully skip 20nm and move right to 14nm production. And even then they would still be two years behind Intel.
The fabrication war is over. Intel has won. The proof is arrival of Intel's 14nm production...