Ask HN: What comes after we hit the 16nm barrier? | Hacker News Reader

Ask HN: What comes after we hit the 16nm barrier?

16 pointssmanek32 comments

For the last fifty years, we have lived in the world of Moore's Law - with truly astounding results.

However, within the next decade we will be approaching fundamental physical limits of the transistor (e.g., http://en.wikipedia.org/wiki/16_nanometer).

What do you think comes next? Will hardware plateau while software catches up? How will this affect the computing future we've been promised (atom-level simulations of complex systems, the singularity, etc.)?

View on HN

Moore's Law isn't going to die. It's going to be extended, (or supplanted) by further developments, like Einstein's gravitation extended Newton's.

There are many, many successor technologies to CMOS out there: nanotubes, nanowires, magnetic switching, molecular switching, photonic switching. I'm sure more have been proposed in the five years since I stopped paying attention.

The situation for semiconductor researchers and manufacturers is a lot like the one Edison faced with the light bulb. We know what we want, and there's a huge field of options open in front of us. Now we just have to spend some time on basic research to figure out what scales and what doesn't. This might take some time, but a delay will only make the next doubling more lucrative when it becomes feasible.

And if you're comfortable with (only slightly) more far-out ideas, the wall is being written right now:

http://metamodern.com/2009/05/22/a-third-revolution-in-dna-n...

It so happens that this development has been anticipated here:

http://e-drexler.com/d/07/00/1204TechnologyRoadmap.html

I think two obvious ways forward are

* Increasing the die size (feasable by reducing substrate costs)

* Stacking transistors (3D CPU's)

Hopefully they'll (re)invent blood vessels along the way--a.k.a. an effective heat removal system like our brains have.

Deleted comment

They can do more than one thing! Why shouldn't they?

On re-read, I was being snarky. I apologize.

We've been living with Moore's law so long that young people seem to think it's similar to an actual law like gravity.

Fact is, computing has benefitted from some core technologies (ex. lithography) that have lended themselves to order-of-magnitude increases in things like number of transistors per square cm. It's been a happy coincidence, but, inevitably, that ship has about finished its trip and is heading back to the dock.

Here's what will happen (and is happening) when hardware hits that limit: pretty much nothing. People will still use computers for email, browsing the web, watching videos, numerically solving partial differental equations, etc. There will be no clamour to make faster computers because computers are pretty much fast enough for what we want them to do. If they're not fast enough, then tasks can be spread out and handled in a distributed manner.

Oh, also, computer hardware companies will find that they can't get you to upgrade every few years to get the latest and greatest. This is great news, btw, for, say, schools; since they'll be able to invest in computers and expect to keep them long enough to be worth the investment.

According to wikipedia, 16nm will be succeeded by 11 nanometer. http://en.wikipedia.org/wiki/11_nanometer

Massive parallelism. Of course we need parallel algorithms as well and deal with all the concurrency issues. I think there's one or two paradigm shifts in the cards as we transition towards this massive parallelism.

But once transistors can't get any smaller the only way to get more cores will be to have physically larger CPUs. That might work for a while. But, pretty quickly, those larger dies will use too much power and get too hot (not to mention be too physically large to fit in the form factors we want).

At best that will buy us another few years of Moore-level growth. But you can't really have CPUs be more than 5 to 10 times larger than they currently are.

Moore's law isn't actually what's terribly important I think. What we want isn't transistor density but processing power. The human brain has a lot of processing power, uses very little energy and works with massive parallelism. But you may be right to question whether we can smoothly transition to some entirely new hardware paradigm quickly enough to double processing power every 18 months. In the long run I believe we can, but we may hit a temporary roadblock in the next few years.

The human brain isn't all that powerful (hardware wise) ...

The numbers I've seen suggest that it's total 'processing power' is only on the order of a few hundred teraflops - and we already have plenty of clusters bigger than that. I know that floating point operations per second is a crappy metric but consider that a neuron can only spike about 100 times per second (i.e., the brain runs at 100 Hz) - even billions of neurons put together just aren't that powerful in terms of hardware.

Software, not hardware, is already the limiting reagent to 'smart' computers.

Everyone says 'massive parallelism' is the answer. But once transistor density plateaus the only way to add more cores (i.e., add more parallelism) will be to make CPUs physically bigger (or make each core 'weaker'). Making each core 'weaker' adds very little net benefit (2 * 100 = 20 * 10) and having physically larger CPUs isn't really practical beyond a point ...

The neuron runs at just 100 Hz, yes, but the brain is not a synchronously clocked machine. How much information processing capacity is due to the timing of spikes? How fine grained is its time sensitivity? 1ms? 1ns? 1µs? 1ps? ...

If it is as you say, all the better. It means that increased processing power could be achieved by making better use of the parallel hardware we already have. Still, I contend that using parallelism is key in some shape or form.

I cannot say much about transistor density issues as I'm not competent in that area. I read that IBM is experimenting with nano technology in order to create much more dense hardware that uses very little energy. I have no idea how far out that sort of approach is.

I mostly agree with you, but I don't think an increase in cpu size will slow the pace down much; it will just shift the cpu load to the server side.

I agree. There seems to be no apparent technology mature enough to sustain hardware complexity growth. I think a pause would benefit software technology, in some sense allowing it to catch up on its potential.

.. you can't really have CPUs be more than 5 to 10 times larger than they currently are

Why not?

And if you can, it doesn't have to be located in your device considering improvements in network technology.

Well, basic physics -- it already takes nontrivial time for signals to cross the die. If you make the die bigger, you're probably just going to be using that to add more cores, and the time for the furthest-apart cores to communicate is going to grow.

Transmission delays suck, but this is something that the brain has evolved to cope with (consider the speed of an action potential vs the speed of electrons in a CPU and the distances between neurons in a typical mammalian brain vs the distances between transistors in a CPU). So I think there there are strategies we could use to deal with larger silicon, but I do agree that there will always be an incentive to keep things as small as possible.

have you heard of amdahl's law (en.wikipedia.org/wiki/Amdahl's_law) before? The push towards parallelism and multicore is a ploy for processor sales. Most applications can't benefit from a parallelism on more than 4-8 cores. There are applications such as computer graphics and servers that benefit, but your average user only uses one core right now. The problem that processors are facing is that clock speed can't be increased. The heat can't be dissipated fast enough and they would melt if run faster.

Most applications can't benefit from a parallelism on more than 4-8 cores.

But since the vast majority of people do not run one or two applications per computer, that doesn't matter that much. I'll go out on a limb and predict that the day is coming when no one ever closes an application. Actually, my parents are already there on their Mac Mini -- they only close windows, never applications, until a software update requires a reboot.

I have indeed heard of Amdahls law, but his function includes a measure of logical dependency. That's what limits parallelism. The dependency between computing my risk for Parkinson's from my genes and computing yours from your genes is exactly 0.

I believe what we should really be interested in is running new types of applications, not speeding up word processing or other single user interactive tasks.

http://en.wikipedia.org/wiki/Reversible_computing Reversible Computing could drastically lower heat output. Of course, it would probably have many effects more profound than that, as well.

It's already difficult for hardware designers to keep up with the underlying technology. It's incredibly difficult to manage the complexity of design, hence the massive growth in EDA tools over the last 10 years.

The productivity gap between the underlying hardware and the capabilities of the designers is probably a more immediate threat than reaching the fundamental physical limits.

Personally I don't care that much. Current processors are plenty fast as it is. I mean, it's always nice to have faster ones and I don't think there will ever be a point where everyone is satisfied and we stop improving them, but I'm just not CRAVING for more speed so much.

However, what I'd LOVE to have is ubiquitous availability of very fast persistent storage. It's really a pain to deal with the relative slowness of hard disks.

Something like memristors would be a solution. http://en.wikipedia.org/wiki/Memristor

I think this question has already been answered -- Moore's Law has been out of play (for processor speed) for some time now. Microprocessor companies are now banking on massive parallelism.

One of the most promising new technologies (imo) relies on probabilistic transistors -- which occurs as they get smaller and more current leakage occurs. A perfect example is video decompression, which does not have to be exact. You can get drastic power reduction by doing it probabilistically (for some applications), which is great for mobile devices.

Moore's law doesn't have to do with processor speed. It has to do with the number of transistors that can be placed on a chip.

From http://en.wikipedia.org/wiki/Moore%27s_law

Moore's law describes a long-term trend in the history of computing hardware. Since the invention of the integrated circuit in 1958, the number of transistors that can be placed inexpensively on an integrated circuit has increased exponentially, doubling approximately every two years.

Speed gains have been a side effect of this.

Yes, I was referring to the fact that Moore's law is still applicable in terms of computer memory, but not microprocessors ... In any regard, I listened to talks four years ago that convincingly stated that it was no longer applicable, so it seems that we've already seen what happens once Moore's Law is no longer accurate.

http://en.wikipedia.org/wiki/11_nanometer

It goes to 11.

My personal belief is that non-semiconductor technologies will come around, at least for specialized tasks. I imagine I'll be using similar technologies we have now for some time longer when it comes to day-to-day operations.

We software developers will have to actually start paying attention to code bloat and performance. We won't be able to count on beefier hardware to run our next round of sloppy hacks!

Optical computing would be pretty cool.

Wouldn't pure functional programming fit that specific bill perfectly?

Reply on news.ycombinator.com