Intel has confirmed a manufacturing process shrink beyond the forthcoming 22nm process due to be used in Ivy Bridge CPUs.
Intel CEO Paul Otellini has revealed that Intel is currently tooling up fabrication plants for 14nm transistor technology, indicating that Moore’s Law is set to continue well past the 22nm 3D Tri-Gate transistors expected next year.
Moore’s Law is a key factor to Intel’s design philosophy, with the Otellini saying that the progression of Moore’s Law is akin to human innovation and progress. It’s no surprise that Intel is always keen to Gordon Moore’s observation proved accurate.
There have been may obstacles to this progress, but ‘
each time, Intel engineers have found a way to innovate past, around and through perceived obstacles using new materials, inventing new technologies along the way,’ according to Otellini.
While the most recent example of Intel keeping Moore’s Law alive is the Tri-Gate transistor, which Otellini described as enabling ‘
new levels of performance and power efficiency across the computing spectrum. The world needs Moore’s Law to continue, and Intel is committed to make this happen.
‘To that end I can tell you that we already have line-of-sight to our 14nm technology. In fact, we are well into development of this technology and are beginning to build and tool our factories to support it.’
Otellini gave to timeline as to when these fabs would deliver the technology for a retail product, not how many (or even which) fabs were being tooled up for the new manufacturing process. However, it’s clear that Intel is doing its upmost to keep transistors shrinking and thus deliver ever smaller circuits for smaller, sleeker devices.
Annoyed that you’ll have to wait too long for the new transistors, or just relived that progress doesn’t show any signs of slowing? Let us know in
the forum.
15 Comments
Discuss in the forums ReplyI found an old PC at work the other day that was hidden in the corner of the IT department (2 desks in the corner) and booted it up to find Win2K and a 130nm Athlon XP. Its amazing how far we have come with transistor technology.
Even at 32nm, as I recall, aren't there Nvidia GPUs with just over a billion already? So, if you don't mind a little scratching on the way down, say, 2.25 bn?
Assuming you dont want to risk choking on your bite of silicon goodness I would assume you would want a bite of around a 2cm sphere. Therefore giving you a bit more at 2.99 billion. Let me know how that works out or you please.
How low can they theortically go?
As it is we are rapidly getting to the point where transistors litterally can't get smaller because they are single atoms. Now really that is a decade + away even if the current pace on miniturization keeps up, but we are getting there in my life time if we don't hit some kind of quantum barrier before then.
I think at some point, I don't know when, but I'd guess in the 6-10nm range making the transistors smaller is going to produce almost no benifit and that we are going to need to focus more on making them switch faster and more efficient. Could be when hybrid electrooptical circuits are needed, or even all optical.
http://img828.imageshack.us/img828/1788/atomroadmap575px.jpg
based on the scientific articles about very early research on carbon nanotube transistors, if that technology were to be made workable it would be near electrically perfect and about 3nm in width.
so quite small, but almost there!
Switching voltages for Transistors are effectively lower than they shoud be in theory
Remember we're talking structure width here...structure thickness has reached the level of one-atom-less-and-it-won't-work-anymore a couple of years ago ;)
Gate oxide thickness has been under 10A for ...gee, half a decade at least? :D