Below 5 nm, can Moore's Law continue?

[N-A-N-0]

http://www.eetimes.com/document.asp?doc_id=1319330

To a noob, can materials like graphene take off? Is quantum computing, at least for the average consumer a realistic goal?

Can chip features go anywhere below 5 nm? Even 5 nm apparently could be quite a struggle and isn't relatively speaking much of a leap over 7 nm. To quote Intel's former chief architect, Robert Colwell:

"For planning horizons, I pick 2020 as the earliest date we could call it dead," said Robert Colwell, who seeks follow-on technologies as director of the microsystems group at the Defense Advanced Research Projects Agency. "You could talk me into 2022, but whether it will come at 7 or 5nm, it's a big deal," said the engineer who once managed a Pentium-class processor design at Intel.
"I don't expect to see another 3,500x increase in electronics -- maybe 50x in the next 30 years," he said. Unfortunately, "I don't think the world's going to give us a lot of extra money for 10 percent [annual] benefit increases," he told an audience of processor designers.




Colwell poured cold water on blind faith that engineers will find another exponential growth curve to replace Moore's Law. " We will make a bunch of incremental tweaks, but you can't fix the loss of an exponential," he said.


DARPA tracks a list of as many as 30 possible alternatives to the CMOS technology that has been the workhorse of Moore's Law. "My personal take is there are two or three promising ones and they are not very promising," he said.

Can transistor densities continue to double beyond the early 2020's (forecasted) and is there any other way to continue to push raw power at the costs and rates of change we've become accustomed to?

 

[jpiniero]

Still think it's dead at 14. That doesn't mean 10 and 7 won't happen, it just will take longer for them to make economic sense over the previous node.

 

[slugg]

One aspect of this "performance barrier" that seems to be ignored often is the software. We still haven't proven whether or not P = NP, so it's still unknown whether or not we can drastically increase the performance of our software. If someone actually proves that P does not equal NP, then we know for a fact that we're forever limited by computers as we know them today. If we prove that P does equal NP, then we know there's a way to speed up our slow algorithms and continue to get exponential performance gains, although through a different way. Nobody has successfully proven either one.

Personally, as a computer scientist (albeit not a good one ), I believe we will never develop a deterministic Turing machine capable of solving NP-complete problems in polynomial time. In other words, I personally think that we are screwed and will in fact hit a performance wall. I have no proof - just a hunch.

 

[Smartazz]

We've already seen electromechanical, relays, vacuum tubes, transistors and integrated circuits used for computing. I have trouble believing that integrated circuits are the end. We may not get below 5nm but maybe someone will figure something out. If integrated circuits got us this far I am afraid to see what the next paradigm will bring.

 

[NostaSeronx]

Moore's Law will continue scientists/researchers will find A WAY!

 

[videogames101]

Vast amounts of time and money are being committed to furthering the research of advanced transistor technologies that can continue to scale in performance. I doubt that much time and money is being completely wasted.

 

[Idontcare]

The end of Moore's law and the end of scaling are two separate topics.

Scaling is simply something we have done as a means to reduce costs.

Reducing costs is what Moore's law is about.

Scaling may be dead, but Moore's law can continue so long as engineers continue finding ways to reduce cost.

So the question is: which conversation do you want to have? Do you want to talk about scaling, and its limits, or do you want to talk about Moore's law and its limits?

 

[AtenRa]

I believe that Scaling will go bellow 5nm and cost reduction will continue (450mm wafers for example).
I just believe that less and less firms will use cutting edge node process going forward. They will stick to one node more than 24months and use the next node only for very low volume high end products.

 

[Atreidin]

Moore's law is about increasing numbers of transistors in an IC. That's it. You can't move the goalposts around by changing the definition in order to make yourself feel better. It isn't about performance, cost, scaling, or anything else. People romanticize it a lot and make alternate "laws" but none of them are Moore's law, which in itself is just an observation, not some universal truth. If you want to measure something else, don't call it "Moore's law", because that isn't what it is.

 

[Borealis7]

this again?...fine i'll chip in.
my opinion is that we're going to see more 3D chip stacking in the future. planar density may stagnate, but density would effectively double if you stack two pieces of silicon on top of each other, like SAMSUNG is doing now with their new Flash technology.

 

[UaVaj]

increase compute power with smaller node.
end of the road not much further. unless alien technology is kept in secret at area 51.

increase compute power with better software to take advantage of parallel processing.
possibility. imagine computing with dual 4960k setup with all 12 core peg at 99% and all 12 ht peg at 99%.

 

[Torn Mind]

Moore's law is about increasing numbers of transistors in an IC. That's it. You can't move the goalposts around by changing the definition in order to make yourself feel better. It isn't about performance, cost, scaling, or anything else. People romanticize it a lot and make alternate "laws" but none of them are Moore's law, which in itself is just an observation, not some universal truth. If you want to measure something else, don't call it "Moore's law", because that isn't what it is.

What Moore actually discussed in his paper, density at a minimum cost per transistor was the topic.

 

[OverVolt]

Vast amounts of time and money are being committed to furthering the research of advanced transistor technologies that can continue to scale in performance. I doubt that much time and money is being completely wasted.

Yea it is :awe:

I don't know why people have blind faith that science can break through physical limitations of the universe, although I'm sure the scientists don't mind taking the money to bang their heads against the wall.

We live in the electric sphere of the universe. When you exert force on something, you distort the electron shell and it pushes back. When you feel something, its the repulsion of the electron shells. All the chemicals and drugs we use, are so useful because of their specific electron shell configurations. In computers we control the flow of electrons. Guess what, the insulator is already only 1.5 atoms thick.

It used to take 6 elements of the periodic table to make a silicon transistor CPU and now it takes over half of them. Once you divy up all the elements cut into the silicon and insulators etc. there isn't enough room at the atomic level for any smaller shit. Get over it. Talking about quantum computers and DNA computers etc. is talking about an entirely new technology its not just a smaller fab, but starting all over again as if you were starting over with the transistor radio in 1950 and zero software. Carbon nanotubes have the potential to make CPUS MAYBE one fab size smaller than topped out silicon, and thats IT.

Listen to what the guy in the article says, he IS the scientist whom you guys have so much faith in cracking the physical limits of the universe.

 

[Abwx]

There s no way scaling will keep on going , unless Sidney Coleman
possibility of the universe collapsing through vaccum decay is right ,
then this will result in another universe with new particles and physics
laws that would perhaps allow for a new form of matter to follow
Bose statistics laws rather than Fermi-Dirac ones , hence allowing
several "atoms" to be at the exact same place simultaneously
and resolving by the way the size limitations but given the necessary
conditions for such a possibility to materialize we can guess that we
wont no more be here either to benefit from it or even just be witnesses...

 

[SlowSpyder]

I remember when I was really just starting to get into the enthusiast side of computers in the mid 90's. Reading up on the technology and learning about how Intel produced CPU's. The end of scaling seemed so far away.

 

[utmode]

Yea it is :awe:

I don't know why people have blind faith that science can break through physical limitations of the universe, although I'm sure the scientists don't mind taking the money to bang their heads against the wall.

+1. When I was in uni studying physical science and that was long time ago. My teacher told me Cambridge trying to make HDD based on electron spin for last 10 years. Anybody seen any news on that. nope. Because its very hard to control tiny element and its all probability at that scale.

I think we need better compiler and software.

 

[krumme]

Intel is starting to reduce capex. Imho It looks like the business forecast tells newer nodes will perhaps slow down. I think its sign we can see the start of the end for the era of Moores law.

Moores law is not a law of physics that works indefinitily within a certain perspective. Like many economic "laws" it works within a certain context typically changing over time. Many seems to forget that. It doesnt mean the law is and was wrong just that the situation have changed.

 

[soccerballtux]

Intel is starting to reduce capex. Imho It looks like the business forecast tells newer nodes will perhaps slow down. I think its sign we can see the start of the end for the era of Moores law.

Moores law is not a law of physics that works indefinitily within a certain perspective. Like many economic "laws" it works within a certain context typically changing over time. Many seems to forget that. It doesnt mean the law is and was wrong just that the situation have changed.

decreasing capex could be because decreasing revenue or operating margin or profit or whatever. IE symptom not "we foresee in 5 years..."

 

[Kippa]

Personally I think that layered 3d circuits are the way to go, stacking layer on layer.

 

[Ajay]

As the end approaches, "when Moore's Law stops it will be economics that stops it, not physics, so keep your eye on the money," he said.

So, Colwell is talking about the doubling of xtors per chip every couple of years at a reasonable cost. He's saying that that exponential relationship is going to end pretty soon - e.g., Moore's 'law' is coming to an end. He also said, very importantly, that the world isn't going to want to pay much for annual 10% gains - something we've already been griping about (IMHO, he is saying Intel, and other semiconductor companies business models are going to come to a screeching halt).

Given his access to what is going on in the semiconductor business because of the position he holds at DARPA, there is a very high probability that that his prediction is correct.

Bottom line, the party is coming to an end in 'CPUs and Overclocking' :\
Pretty soon it's going to be BYOB and no free snacks.

 

[Smartazz]

I think the most important metric is operations per second per constant dollar. I don't know why this is rarely talked about.

 

[OverVolt]

You gotta understand that intels fabs keep increasing in cost. Its really outpacing the gains at this point.

With Moore's law, computers always got smaller, faster, and cheaper.

Without Moore's law computers are: Smaller, Faster, or Cheaper

-Pick Two

You already see this developing. There are still big clunky desktops, they are all fast these days, they are cheap, but they aren't getting smaller.

You have tablets, smart phones, and netbooks. They are smaller and cheaper, but not fast.

You can have fast and small, but it won't be cheap.

 

[Fjodor2001]

Unfortunately, "I don't think the world's going to give us a lot of extra money for 10 percent [annual] benefit increases," he told an audience of processor designers.

So what's he saying? The semiconductor industry is about to die a slow and painful death? :'(

 

[TuxDave]

When talking about the end of Moore's Law, I'm always reminded of this speech when I attended ISSCC.

http://www.intel.com/pressroom/archive/speeches/moore20030210.htm

Well, many of the parameters related to the semiconductor and solid-state circuits industry have shown exponential dependences over the years. But no physical quantity can continue to change exponentially forever. There's always some kind of catastrophe if you project it far enough into the future.

What I want to do today is look at some of these exponentials and maybe give some idea where they might go and talk a little bit about how we're going to deal with the looming catastrophes that people seem to be projecting as they look further down the road.

 

[Idontcare]

So what's he saying? The semiconductor industry is about to die a slow and painful death? :'(

Not at all. Well, yes, he is quoted as saying something that can easily be interpreted as that but I really doubt that was the context for his statement.

Look at space travel versus automobile.

Both went through their meteoric rise in popularity and CAGR. Space race climaxed and most certainly crashed after the Apollo program was cancelled.

Automobiles don't give us 10% per year improvements and yet the industry is thriving. (we just don't treat them like disposable assets )

Now some people want us to view the ending of Moore's Law as being akin to the last chapter of the space race to the moon, after which everything is supposedly going to implode into a singularity and the semiconductor industry will shrivel into a shadow of its former self as the money stops flowing, consumers stop spending, and innovation stops happening.

That was, to a large part, how the space race fizzled. No moon bases, no trip to Mars or Jupiter's moons, no space hotels, etc.

But that was because people were not dependent on that aspect of space travel. We became critically dependent on satellites though, and the commercial satellite business has steadily grown year after year, albeit off the front page of any major newspaper.

And look at the auto industry, we are critically dependent on our vehicles. So despite being a mature industry, and despite the rather abysmal rate of improvements in terms of energy efficiency and safety, the industry continues to grow larger every year (recession induced dips withstanding).

So which do you really think the semiconductor industry is going to follow? The space race or the auto industry?

I think we are too dependent on our computers, for both work and play, to decide we aren't going to pay for a 10% increase every year. I think we will continue to pay for it, but we won't be replacing stuff every 12 months anymore.