Replies to post #37775 on Intel Corp (INTC)

[Sarmad]

>> Diesize-economies still exist, it's the parametrical returns that diminish currently and will finally go away. Not the question of if, but when.

Yes, the question is when ? Is that next year (at 45 nm).
is it in 4 years (32 nm) ?
in 9 yrs (22 nm) ?

The way I see the current situation, AMD's entire problem is that they are behind on node and process technology. That suggests that technology competence is not equally distributed between Intel and AMD.

Also, my observer-understanding is that some key advantage will come to whoever gets to 45 nm first. Basically, integrating graphics with processor becomes feasible at that node. If AMD's retardation increases to 2 years (from the current 1.2 years), wouldn't that be a key disadvantage ? maybe fatal ?

[alan81]

Regarding parametric process improvement...
it's the parametrical returns that diminish currently and will finally go away.
It is important to understand why this is and what is being done about it. The problem is the gate oxide has not been scaled for the past several generations because they ARE finally at a physical limit. As such, the last few processes have returned less than historical improvements. However, with the recent rumor that INTC is going to a new gate oxide material opens up the opportunity to get back on the scaling train. Depending on the quality of this new solution the amount of improvement we will see can vary significantly. At the best case for Intel, they not only return to the historical improvement rate, but they return to the performance one would have extrapolated from the process nodes prior to the gate oxide scaling reduction. At the other end of the spectrum it allows them to simply continue at this 15% or so improvement per node rate, from where they are.
--Alan