Wednesday, April 23rd 2008

TSMC Outlines New High-k/Metal-gate 32nm Process

Responding to IBM and its development partners that took the lead in the nm battle by announcing the availability of a high-k and metal-gate 32nm offering, Taiwan Semiconductor Manufacturing (TSMC) today outlined more details about its 32nm process. Rick Tsai, president and chief executive of the company said "We will have high-k and metal-gate at 32nm," during a keynote at the 2008 Technology Symposium. Jack Sun, vice president of research and development, also indicated that TSMC will develop a third-generation triple-gate oxide technology for its 32nm low-power process. For its 32nm high-performance process, TSMC will offer a high-k and metal gate technology, Sun added.Source: EETimes
Add your own comment

8 Comments on TSMC Outlines New High-k/Metal-gate 32nm Process

#1
a111087
haha, that good for amd
Posted on Reply
#2
Silverel
Eh, I'd imagine they have to license the HKMG process from Intel. Something along the lines of copyrights and such, if indeed Intel owns the rights to it.

Doesn't TSMC make ATI and nVidia cards? Imagine GPU's with the OC power of a C2D... hehe
Posted on Reply
#3
magibeg
Shrinking them doesn't always mean that they will be able to OC higher. Makes them more sensitive to voltage with the smaller parts and such but should hopefully reduce heat and power consumption with a minimal performance increase. Also means cheaper to make :D
Posted on Reply
#4
Silverel
The HKMG process eliminates a whole ton of voltage leaking, making them less sensitive as the voltage stays where it is supposed to be. It's not the only reason that Intel can OC so well, but its still a big one. I think Anandtech had an article on the C2D line when they came out explaining all the technical stuff like that.
Posted on Reply
#5
R_1
Good for TSMC, but I wonder why ATI and NVIDIA do not use 45nm process form 2007 or 40nm process announced this year by TSMC. Instead of that they ordered GPU's on 65nm and 55nm(Ati).
Posted on Reply
#6
Steevo
A die shrink requires a whole new revsion to be made on the core. And tech that is in production is to far along to back out on.
Posted on Reply
#7
iLLz
Hopefully this will help push GPUs into uncharted territory, so long as AMD and Nvidia order new GPUs utilizing this tech. Soon enough (I hope), Intel's Larabee will hit the market and then there will be 3 Major players in the discreet graphics card business. Competition is good for us consumers for sure!
Posted on Reply
#8
phanbuey
by: R_1
Good for TSMC, but I wonder why ATI and NVIDIA do not use 45nm process form 2007 or 40nm process announced this year by TSMC. Instead of that they ordered GPU's on 65nm and 55nm(Ati).
Designs that work on 65nm and 55nm may not function at all at 40nm and 45nm.
Posted on Reply