Intel to present 32nm chip while AMD shows off 22nm part
Duel of the memory cells
Intel will release details of its 32nm chip fabrication technology in just over a month's time. But AMD will be on hand to talk up its 22nm process.
Representatives from the chip giant will be presenting a paper at the IEEE International Electron Devices Meeting (IEDM), which kicks off in San Francisco on 15 December. They will discuss how they produced a 291Mb SRam memory array to test the process, which uses high-k and metal gate technologies. It has a density of 0.171μm², EE Times reports.
But we can add that Intel won't be the only chip maker detailing its progress to the 32nm node. Semiconductor foundry TSMC will be present to discuss its 32nm 2Mb SRam chip, while IBM boffins will be presenting its 0.157μm² 32nm SRam cell.
AMD and IBM's smallest functional SRam cell up close
IBM will also be partnering with AMD to present what they claim is the "smallest functional SRam cell ever made" - a 22nm high-k and metal gate part with a cell density of 0.1μm²
Static Ram cells are the structures chip makers always build first when testing a new fabrication process.
Intel has already said its first 32nm processors, codenamed 'Westmere', will debut in late 2009. They will be die-shrink versions of its new microarchitecture, 'Nehalem', which is due to debut later this quarter.
COMMENTS
@a square micrometres is an area not a density
It's implied - the area taken up are in units of area (\mu m)^2 ( or (\mu m)^2 if we're being annoyingly pedantic - it's not ambiguous at all...!), therefore if 1 things takes up area A its density [this thing being assumed to be 2d, not 3d] is 1/A (units of \mu m^{-2} ). In conclusion the measurement in square-micrometers carries across directly to the 2d packing density.
It's all a matter of what works - nobody started moaning on when all the Higgs related articles were saying the Higgs mass is expected to be ~ 130GeV. GeV is energy, but since this relates to mass via E=mc^2 we just write down the energy. Particularly useful when the thing's moving, since we need to separate the rest mass [energy] from the energy it has by virtue of its motion, excited degrees of freedom, etc.
@F Seller
"I don't think "micro-square-meter" are official or even used anywhere"
It's horribly ambiguous, so I don't think anyone *should* ever do it. As it happens, I think 1μm² *ought* to be interpreted as 1μ(m²) because that's the precedence of the mathematical operations involved, and the nice thing about SI units is that they *are* mathematically well-behaved and so anyone who does anything to break that is a total plonker. That said, whenever I meet such things in real like I always apply sanity checks and it is usually obvious from the context that they are indeed plonkers.
For a related example, consider the "hectare", which is apparently 100 of some utter abomination called an "are" which is 100 square metres and (allegedly) the "official" derived unit of area in the metric system. Does that strike you as a remotely rational unit of area for a system that measures lengths in metres. Thought not.
Now can we get back to sensible units like nano-furlongs?
@brainwrong
1μm² means 1e-12m², so the () are implicit. "Square-micrometer" - it don't think "micro-square-meter" are official or even used anywhere.
a square micrometres is an area not a density
The title says it all... describing square micrometers as a density is like calling litres a weight.
I guess describing units of measure accurately is just another thing that's become unimportant...
