End of Moore's Law in 10-15 years?

javipas writes "In 1965 Gordon Moore — Intel's co-founder — predicted that the number of transistors on integrated circuits would double every two years. Moore's Law has been with us for over 40 years, but it seems that the limits of microelectronics are now not that far from us. Moore has predicted the end of his own law in 10 to 15 years, but he predicted that end before, and failed."

Comments

it's the law

[User 956]

Moore has predicted the end of his own law in 10 to 15 years, but he predicted that end before, and failed.

So then it seems with regards to his Law, Moore has fallen prey to Murphy.

Re:it's the law

[click2005]

What about the inverse of Moore's Law.. Every 2 years, the average IQ of all users on the internet halves.

Re:it's the law

[Mr. Sketch]

Every 2 years, the average IQ of all users on the internet halves. So true. I believe we have now coined the click2005 law.

Re:it's the law

[Anonymous Coward]

I believe we have now coined the click2005 law

"we"? "we" didn't do shit, n00b.

Re:it's the law

[tomstdenis]

But the IQ is the average ... so it can't halve. :-)

Re:it's the law

[somersault]

It can if you count all the people that don't use the internet? As long as they gain in current IQ terms sufficiently to balance it all out :P

Re:it's the law

[jellomizer]

I would agree the Average IQ of the people on the internet is lower but it is not like moore's law inverse it is aproaching a point where it is matching the worlds average IQ

Early on the IQ of people who used the Internet was much higher then the Average General Population IQ because inorder to use the Internet you normally needed to be in College and/or have attened college.
Then it dropped to the people that could afford the high price of the Internet with the cost of the computer.

Now as the internet is getting faster and cheaper as well as the computers more average and below average people are having resources to access the internet so the average Internet IQ is dropping.

Re:it's the law

[fbjon]

What about the inverse of Moore's Law.. Every 2 years, the average IQ of all users on the internet halves. That's related to Godwin's law. In fact, I'll make a prediction:

In 10-15 years, we will see the End of Godwin's Law, as internet forums reach their theoretical limit of inanity and flaming. Unfortunately, this prediction will also fail.

Re:it's the law

[PingPongBoy]

but he predicted that end before, and failed

And so we should have Moore's Sliding Window: Moore's Law to fail in 10-15 years, renewable annually.

Vegas might want to give odds on Moore's Law failing after a certain time, payable well after the fact.

It does get harder and harder to uphold the law. AMD is in dire straits, as its new quad core is failing to turn enough heads while a massive amount of debt needs to be repaid. If consumers care to keep Intel pricing at bay and to encourage more research, they should buy some AMD-equipped computers. The price level of a "good-enough" computer keeps falling though. Software isn't doing anything that makes people want faster hardware immediately, other than adding bloat. Even web pages are so loaded with junk that they take too long to load--there should be a "junk off" option, and web developers should delimit what is the interesting part and what is the crap so that browsers can slice a page apart. The situation seems so counterintuitive and pathetic.

So what might really be the limit to Moore's Law? Consider the human brain. It consumes less power than a CPU but can do amazing thinking, although not long, quick sequences of arithmetic. A billion-core CPU might easily be achieved in the near future. The cores wouldn't be too impressive by themselves but still individually capable of long, quick sequences of arithmetic, and put all together can rival a human brain. Where's the software? Is that what's really holding back hardware development?

CPU design is somewhat easier than software design, I suppose, since a CPU just has to achieve a general-purpose functionality. Software has to be better with each generation, and suit specific desires, while following longer and longer pathways.

Re:it's the law

[corifornia2]

I think Moore and Edward Murphy should have a bikini clad mud wrestling battle to determine the Ultimate Law to be Referenced on Slashdot.

Re:it's the law

[tomstdenis]

But people not using the net will be the minority and not influence the average. See where I'm going with this?

You should have said "with reference to pre-internet era testing criteria," because you can't even compare a 1980 IQ to a 2020 IQ since again, they're both average and there isn't a "score" otherwise.

Re:it's the law

[Jarjarthejedi]

Normally that holds true, but this is the internet we're talking about...the average IQ can decrease relative to itself because people can be just that stupid. Not even mathematics can keep up with the drop in the internet's IQ.

Re:it's the law

[somersault]

You could get someone to take both tests and I'm sure their score won't be the same. If they score lower then you know that the average is increase, and vice versa (or you may decide that the tests are a load of bollocks, which is probably the more adequate explanation). Maybe I'm underestimating the percentage of the world that uses the internet, but you could also be overestimating it. I don't know the composition or size of the groups that they decide this 'average' from anyway, but just because the average intelligence of the world changes doesn't mean that the original test will change. Though you can always re-weight the answers to the questions.

Re:it's the law

[TrekkieGod]

It can if you count all the people that don't use the internet?

Damn! You've just stumbled upon the Amish plan to take over the world. By not being on the internet while the rest of us dumb down, we'll soon be completely dependent on their superior intelligence!

Re:it's the law

[Marxist Hacker 42]

What about the inverse of Moore's Law.. Every 2 years, the average IQ of all users on the internet halves.

Which of course raises the question- what will the average consumer do with all of this? When your phone is more powerful than the desktop, there seems to be a natural endpoint like what happened with word processing- a point where the used market is larger than the new market, and everybody can afford enough technology to do anything they reasonably want to do. I suggest that point will be reached *before* the endpoint of Moore's law- and thus the law of supply and demand will insure that we NEVER reach the endpoint of Moore's law, because there will be no demand to do so.

Re:it's the law

[zippthorne]

there should be a "junk off" option, and web developers should delimit what is the interesting part and what is the crap so that browsers can slice a page apart

There already is such an option. It's called javascript. Web developers always put all the crap behind a javascript wrapper. Occasionally they also use the "Flash delimiter" but it's often behind a javascript wrapper too, for good measure.

Re:it's the law

[DavidJSimpson]

I believe that the law to which you refer is the Law of Conservation of Internet Intelliegence, which states: "The sum of the intelligence of all Internet users is constant." The collary of this law is that as the number of users increases, the average intelligence of each user decreases.

Unfortunately, I am one of those Internet users, and therefore my own intelligence is decreasing as more of you join the Internet. Thanks.

Re:it's the law

[kesuki]

the best measurment of the drop in the 'internet IQ' is the dollar value of 'cybercrime' as the drop in internet users iq's falls, the dollar value of money stolen from the average internet user rises. a smart preson sees an 'email' from 'paypal' requesting they log in on a url that looks like paypal in the email, but in the url bar at the top shows some hotel in russia and they don't log in. as more and more people who aren't smart enough to recognize simple phishing emails when they get them log in and have their identity and cash stolen the overal cybercrime value increases.

in a less obvious scam someone goes to google looking for cheap software, sees an ad for $40 unlimited software downloads, or a site selling 'download only' software and foolishly gets their credit dard info stolen, by crooks, and pay for pirated software... well i personally am intelligent enough to google the companies name and find them in a database of cybercrooks, but what about the millions who just think the deal is real, and wind up screwed over... well, that's why cybercrime has reached new records...

america has the highest reported rate of computer viruses, simply because finding and using security software that really works is too 'hard' for the average user, they just buy what the guy at the computer warehouse said was good. and some are stupid enough to not even do that. all the 'user friendly' suites have shortcomings that virus programmers and hackers can find workarounds for. because they generallz take a 'less secure' approach to defualt security settings to avoid having tech support overwhelmed with users who can't get their internet game, or websites to load.

and often times those techies at the warehouse stores use 'default' fileshare names, and turn filesharing on, and then hackers can easily find and compromize those systems.

people wind up throwing away the systems that got viruses-trojaned because the cost of a new computer is less than having a computer guy fix it. and usually those systems have to get so bad with malware and viruses that they're unusable, probally they were compromized for years.

hackers get good training on compromizing firewalls, traversing nats, and placing backdoors on vulnerable systems, with all the computers out there... which is why an unprotected windows system takes about 12 minutes to get compromized on the net.

Re:it's the law

[hawk]

That's a corollary of Hawkins' Second Law: "There is no lower limit to human intelligence."

hawk, who has since forgotton his own first law

Re:it's the law

[Tablizer]

What about the inverse of Moore's Law.. Every 2 years, the average IQ of all users on the internet halves.

I was gonna call you a nasty name for saying this, but I forgot what the name was.

Re:it's the law

[not-enough-info]

You can't assume that IQ is a measure limited to a demographic (internet users).
It could very possibly be fact that within the general populace, only high scoring people initially use the internet.
Then, as many low scoring people are added to the internet demo, the average could very well halve.
(150 + 50 + 50 + 50)/4 = 75

If the average IQ is halving, this means that a good percentage of high scoring people are staying off of the internet.

Re:it's the law

[VENONA]

I can't agree. Even if consumers who always have to have the latest (even if they don't really need those high-end specs, and even if you can satiate the gamers (who do use those high0end specs), the technical computing folk will continue to drive for more compute power.

Meanwhile, thinner transistor gate structures, etc., will also improve power efficiency--already the leading cost in many datacenters.

Re:it's the law

[Frozen Void]

You haven't seen the bad parts of the internet.

Re:it's the law

[RockDoctor]

click2005 :
Every 2 years, the average IQ of all users on the internet halves.

Mr. Sketch (111112) :
So true. I believe we have now coined the click2005 law.

No, Mr Sketch, we haven't "coined" the click2005 law, we've witnessed click2005 coin the click2005 law.

But yes, it is a nice demonstration of the postulated click2005 law.

Don't give up the day job. Unless you're a lawyer, in which case DO give up the day job.

Re:it's the law

[somersault]

Why do I get the feeling that some people think "the world" means "America"? :p

Re:it's the law

[Psyjack]

So, would that be Murphey Moore's Law, or Moore Myrphey Laws

Re:it's the law

[jo42]

Digg did a complete bung-up of that rule of thumb. In one year they caused the average IQ of the Internet to be less than 1/100th of what it used to be...

Gordon Moore

[morgan_greywolf]

Is Gordon Moore crying wolf again?

Actually, I don't think it will matter. In 10-15 years, the emphasis will shift away from traditional binary computing and towards quantum computing anyway, making Moore's law sorta moot.

Re:Gordon Moore

[sexybomber]

Actually, I don't think it will matter. In 10-15 years, the emphasis will shift away from traditional binary computing and towards quantum computing anyway, making Moore's law sorta moot.

And if quantum computing should herald The Singularity, then it's definitely moot, since no predictions (Moore's Law included) can be made about post-Singularity computing.

Re:Gordon Moore

[plover]

I don't think quantum computing will be the future for general-purpose computing, and certainly not in 10 years. I think you're nearly right in that the future will lie in parallel computing -- increasing the number of CPUs will be the path to higher throughputs (which coincidentally aligns nicely with Intel's goal: sell more CPUs.)

Either way, when Gordon Moore eventually dies he will still be overflowing the (long)money; variable.

Re:Gordon Moore

[oliverthered]

quantum computers aren't really general purpose machines and wouldn't be able to replace traditional CPUs for a lot of tasks.

Re:Gordon Moore

[Anonymous Coward]

One problem in these discussions is that different people use different definitions of "Moore's Law." Strictly the law is an observation about the increasing density of transistors (i.e. decreasing size of each transistor). However, as we all know, many people simply use the term "Moore's Law" loosely, referring to all exponential increases in computing power.

There is no doubt that we will reach a hard physical barrier beyond which we cannot shrink individual transistors any longer. This limit will be reached in a decade or two, and is probably what Moore is referring to: at our current scaling, we will hit atomic limits rather soon.

But, that doesn't mean that the exponential increases in computing power will end. There are many other things that may happen, such as figuring out ways to build microprocessors with transistors stacked in 3D (rather than having a single 2D layer of transistors), which will increase the transistor count in our computers by orders-of-magnitude. Improvements in chip designs, layouts, and algorithms are other areas that may see improvement. Specialized and dynamically re-programmable chips may also provide us with further advances. Or perhaps, as you pointed out, quantum computers will become viable and mainstream.

There is no guarantee that these more exotic technologies will work out. Yet the microelectronics industry has surprised us time and again with their ability to overcome huge technical obstacles. Thus it seems at least possible that they will deliver technology that is very much up against the physical limits of what is achievable. And with regard to those physical limits, the hard-wall the Moore is predicting in 10 years is only one aspect. There are many other ways for this technology to be advanced.

Re:Gordon Moore

[Velveeta_512]

I was thinking the same thing, or if quantum hasn't been quite refined as a useful science yet, at the very least, nanocomputing should be advanced enough to the point that it can take the reigns from microcomputing... That's what we like to do in this industry anyway, just change out prefixes when the technology hits a certain milestone... If you think the nano-itx motherboards are small, wait until you see the super-mega-ultra-nano-itx in 10 years, all components will be on separate cards, not quite risers, because they'll be connected via fiberoptic links directly into the bus a few microns away from the tri-tri-core CPU, the AMD Zelda-core 64000+ processor... And thanks to the miracle of nanocomputing's molecular level manipulation, food replicators will be a USB device available from Newegg... However, thanks to the miracles of the evolving software industry to meet the ever-increasing needs of users, there won't be drivers available for the new food replicators on Linux, it takes Windows 2020 over an hour to produce 1 hardboiled egg that you could have made yourself in 5-10 minutes, because it's thrashing your solid state drive (because really people, 2 terabytes of RAM is OS's recommended *minimum*), and it tastes more like a peanut butter sandwich because you don't have the latest Microsoft version of the driver for Windows 2020, because of course it makes the hardware incompatible with Windows 2020... Mac users will complain that they've had their iReplicator for 6 months already and stood in line for 12 hours to get one, and Jobs is slashing the price in half in anticipation of the release of the 2nd generation iReplicator...

Re:Gordon Moore

[kebes]

A realistic design for a quantum computer would probably have a classical CPU that does most of the work, with a quantum co-processor. Traditional things, like running the OS and dealing with hardware I/O, would probably still be classical. The quantum co-processor would be assigned computations by the CPU that can be accomplished much faster than on the classical CPU.

This abstraction would mean that most software wouldn't have to be written with any understanding of quantum computing: libraries and compilers would be designed to use CPU calls that launch the quantum co-processor, if available.

For many operations, the quantum CPU would not be needed. But for certain tasks, it would provide orders-of-magnitude speed boosts. If quantum co-processors became commonplace, we would see improvements in all kinds of parallel-processing tasks (matrix operations, simulations, graphics, maybe even search?).

Re:Gordon Moore

[drmerope]

I don't think it will happen this way. First, GP computing is just a small slice of the VLSI application space. It isn't obvious that all the other problems in the world will suddenly become quantum computing problems. I see two effects: first because of cost reasons performance gains will not appear across all sections of the GP market. e.g., home systems will begin to fall further behind top-end computing. Second, Moore's law will slow down before it "ends" at a physical limit. So if it predicts we have five more generations to go, those may smear out over 20 years not 10.

Also: there are many other problems in silicon manufacturing not relating to feature size (the primary subject of Moore's law). i.e., there are ways we could build faster (but not denser) devices than we do now. Thus, the density aspect may stall and emphasis could return more to performance. e.g., using photonic interconnects for long distance signaling, lower Vt with less leakage, etc. The emphasis could also shift to cost. If the per-design mask costs go down than these really advanced process technologies will be available to more low-volume application specific situations. That could be potentially a very big deal.

Re:Gordon Moore

[Daniel_Staal]

Somewhere, once a upon a time, I saw an article that took the opposite approach: They worked out what the absolute maximum transistor density was, and worked out from that when Moore's Law had to end. They figured one transisitor per Plank-unit, in a spherical computer. (Where the clock speed is proportional to the size of the sphere, governed by the speed of light.)

IIRC, it ended up something like 150 years in the future.

Re:Gordon Moore

[Daniel_Staal]

Sorry, correction: The clock speed is inversely proportional to the size of the computer. (Smaller computer, faster cycles.)

Re:Gordon Moore

[Surt]

There's hope for transistor density after atomic level transistors, actually, because density is a measure of transistors per unit volume, not per unit area. We currently talk about the density in terms of transistors per surface area of the chip, we really need to be thinking about the density of transistors you can put in the volume currently occupied by a cpu plus heatsink.

There's probably another 30 doublings that could plausibly be accomplished before the true 3 dimensional density is no longer realistically improvable.

Re:Gordon Moore

[SeekerDarksteel]

I agree with what you're saying in general, but I think one thing requires clarification.

Quantum computing will not provide some sort of magical bullet for parallel calculations. Yes, you can do a lot of calculations in parallel, but you can't get the answers because the qubit(s) holding the superposition of all possible outputs decoheres when you measure it into only one possible output. Quantum algorithms that actually do anything interesting are very narrowly focused and rely on complicated things like the quantum fourier transform or manipulating the qubit(s) so that they're more likely to be measured as the answer you want.

Re:Gordon Moore

[kebes]

I'm not sure if this is the same article that you saw previously, but this paper discusses that topic:
Seth Lloyd, "Ultimate physical limits to computation" Nature 406, 1047-1054 (31 August 2000) | doi: 10.1038/35023282 (for those without access to Nature articles, this arXiv preprint appears to be the same article).

The article reviews the absolute maximum limits for computation, based on current understanding of thermodynamics, relativity, and quantum mechanics.

The basic conclusion of the paper is that a theoretical 1 kg computer (confined to a volume of 1 liter), operating perfectly at the edge of what is physically possible could compute 10^51 operations/second on 10^31 bits of information (as compared to our current computers: 10^10 operations/second on 10^10 bits). Naively scaling Moore's law from current sizes, this suggests that we will reach such limits in 250 years. Of course the paper repeatedly points out that this is for an unrealistically 'perfect' computer, that is somehow able to perfectly organize all its internal matter solely for performing the computation at hand. For instance when running a computation it effectively has a temperature of ~10^9 Kelvin, which is considerably hotter than any known material could withstand.

Nevertheless, it's interesting to see what the fundamental principles of relativity and quantum mechanics indicate as a boundary for any sort of computation. The article is an interesting read.

Re:Gordon Moore

[Anonymous Coward]

Can you imagine the linux scheduler debates over using a quantum processor?

Re:Gordon Moore

[Daniel_Staal]

That looks like the article, thanks.

Re:Gordon Moore

[TeknoHog]

Can you imagine the linux scheduler debates over using a quantum processor?

I don't see any debate there, if you can have something like (|CFS> + |SD>)/sqrt(2)

Re:Gordon Moore

[russ1337]

Sounds like this computer approaches what you're talking about...

Re:Gordon Moore

[NoOneInParticular]

I'm not sure if quantum computing can truly exhibit general computation properties better than the digital computer. QC is great for solving some problems that are intractable for digital machines, but computers that are only capable of computing something that with 99.9% probability is the right result will not easily drive the next generation of OS(*).

I personally think there is a world-wide demand for 5 quantum computers, max. [grin]

(*) for OS in [Operating Systems, Office Suites]

Re:Gordon Moore

[xPsi]

For instance when running a computation it effectively has a temperature of ~10^9 Kelvin, which is considerably hotter than any known material could withstand.

The quark gluon plasma created at RHIC (and soon at the LHC) is hovering at about 10^12 Kelvin (granted for about 10^(-23) seconds)...So there's one material anyway...

Nice post. Thanks for that article.

Re:Gordon Moore

[Anonymous Coward]

oh god, someone mod this up.

Re:Gordon Moore

[Rhapsody Scarlet]

Actually, I don't think it will matter. In 10-15 years, the emphasis will shift away from traditional binary computing and towards quantum computing anyway

I was thinking similarly, but more along the lines that it won't matter so much because Gordon Moore will most likely be dead in 10-15 years and won't have to listen to ridicule if he's wrong again.

That's the Nostradamus way of making predictions.

Law?

[haystor]

Can we stop calling a prediction a law?

Re:Law?

[UbuntuDupe]

When a prediction is confirmed over a long enough time and with enough consistency, that makes it a law.

Although at this point, I think it can only be justified as between "hypothesis" and "theory", but I'm no expert.

Re:Law?

[Cutie Pi]

Can we stop being so f'ing pedantic every time a story mentions Moore's Law?

Re:Law?

[Blinocac]

It doesn't really make it a law, but that is what normally happens. The law of gravity, just a theory. The laws of physics, theories really.

Re:Law?

[IndustrialComplex]

Can it even be a theory? There are far too many variables that are dependant on human decisions to treat this as some natural law.

It seems more like "Moore's surprisingly accurate prediction and continued observation" than Moore's law, but until I can figure out a catchy acronym I think we are stuck with Moore's law.

Re:Law?

[plover]

Can we stop being so f'ing pedantic every time a story mentions Moore's Law?

This is slashdot, so ...

no.

Re:Law?

[Larry Lightbulb]

"Moore's SAPACO" perhaps?

Re: Law?

[Anonymous Coward]

Not unless you want to throw out Boyle's Law, Charles' Law, the Law of Universal Gravitation, etc. Sorry, but there's a name for predictions supported by significant evidence but without positing a specific explanation.

http://en.wikipedia.org/wiki/Physical_law

Re:Law?

[nomadic]

Can we stop being so f'ing pedantic every time a story mentions Moore's Law?

No, because that would be breaking Slashdot's Law.

Re:Law?

[mollymoo]

I suggest you look up what "law" and "theory" mean in the context of science.

Re:Law?

[Surt]

I'm sorry, but law is completely correct, and calling it Moore's law complies with definition (1A!) from websters:
http://www.m-w.com/dictionary/law
1 a (1) : a binding custom or practice of a community

Re:Law?

[Phisbut]

No, because that would be breaking Slashdot's Law.

Gaaahhh! It's not a law, it's a theorem. Stop using words when you don't know their meaning!

Re:Law?

[mollymoo]

Can we stop calling a prediction a law?

Not when that prediction is also a law, no. All it takes for something to become a law (in the scientific sense) is a consistent observed pattern. The law of gravity, for example, is nothing more than concise expression of a consistently observed pattern of behaviour. Moore's law is also a concise expression of a consistently observed pattern of behaviour; it is thus a law.

Re:Law?

[SL Baur]

Moore's Law has roughly the same significance as Bode's Law http://www.alcyone.com/max/physics/laws/b.html

The form of Bode's law (that is, a roughly geometric series) is not surprising, considering our theories on the formation of solar systems, but its particular formulation is thought of as coincidental.

Re:Law?

[haystor]

Except that his prediction is about engineering. The "Law" basically states that we will compact transistors into half the space every 18 months.

"Moore's (current) Rate" would be more appropriate than law, theory, hypothesis. We have the speed of light, not "light's law" after all.

Re:Law?

[adavies42]

It's a law in a very old-fashioned scientific sense--it expresses a proportion between two quantities (transistor density and time). This is the same sort of law as Newton's First Law, Boyle's Law, Coulomb's Law, etc. This use of the term is mostly obsolete, so I'm not surprised it's often understood.

Not to worry...

[Marc Desrochers]

It will be just in time for the arrival of cold fusion.

Re:Not to worry...

[Anonymous Coward]

It will be just in time for the arrival of Duke Nukem Forever. There, fixed that for you.

Moore's Second Law

[MyLongNickName]

Moore's second law: "Moore's first law will only work for 10-15 more years".
Moore's third law: "Moore's second law applies from the time it is quoted not from when it was originally uttered".

Re:Moore's Second Law

[irtza]

Wait, I thought the third law was "Moore's second law takes effect 10-15 years before Moore's first law no longer holds true". I hope I get this straight before the test.... there is going to be a test, right?

Re:Moore's Second Law

[rumblin'rabbit]

I thought the second law was:

The number of experts predicting the end of the first law doubles every 18 months.

Things never change

[Anonymous Coward]

It's funny how we hear the same predictions over and over: Moore's law done in 10 years, fusion power in 50 years, Iraq pull out in 18 months, hard AI in 10 years. The date never gets any closer.

Re:Things never change

[pipatron]

Don't forget a much improved battery technology and flexible, outdoor-readable LCD in 3-5 years.

Moore's Law

[Poromenos1]

The complexity for minimum component costs has increased at a rate of roughly a factor of two per year ... Certainly over the short term this rate can be expected to continue, if not to increase. Over the longer term, the rate of increase is a bit more uncertain, although there is no reason to believe it will not remain nearly constant for at least 10 years. That means by 1975, the number of components per integrated circuit for minimum cost will be 65,000. I believe that such a large circuit can be built on a single wafer.

That's from Wikipedia. He actually said that the cost would halve every year.

Re:Moore's Law

[renrutal]

That's from Wikipedia. He actually said that the cost would halve every year.

{{citation needed}}

Again?

[dylan_-]

There are always a few of these.

I do recall someone telling me that no CPU would ever run at more than 2GHz, as it would then start emitting microwave radiation...

Re:Again?

[JoelKatz]

When I was in college, I learned all the reasons the features on a CPU couldn't be significantly less than 1 micron in size. I also learned that 20Kbps was about the theoretical upper limit for modems.

Re:Again?

[krray]

I do recall someone telling me that no CPU would ever run at more than 2GHz, as it would then start emitting microwave radiation...

I remember having / making a similar claim myself way back when -- with the 486/33 and 486/66 being the hot system in the day. I predicted they'd have a hard time getting above ~80Mhz because of FM radio interference / shielding problems. Boy was I wrong.... :*)

Today I predict "Moore's Law" to hold pretty true -- even in 10 or 15 years. IBM has been playing with using atoms as the gate / switch which will make today's CPU's look like Model T's.

In the 90's they had http://www-03.ibm.com/ibm/history/exhibits/vintage/vintage_4506VV1003.html
Not too long ago they've done http://domino.watson.ibm.com/comm/pr.nsf/pages/news.20040909_samm.html
And recently it has been http://www.physorg.com/news107703707.html

This will both be a boom for storage and the chips themselves IMHO (not to mention my stock :).

Re:Again?

[Dolda2000]

I do recall someone telling me that no CPU would ever run at more than 2GHz, as it would then start emitting microwave radiation... Thinking of it, it would make sense. Really, how comes they don't? Are they just sufficiently shielded?

Re:Again?

[TheRaven64]

I also learned that 20Kbps was about the theoretical upper limit for modems This wasn't too far off. I never got much more than this from a MODEM, since I lived in a rural community with quite noisy lines (26.4Kb/s, I think, was the most I got). Of course, now we don't use MODEMs for communication (well, technically we do, but not in the sense that the word is commonly used).

Re:Again?

[dylan_-]

Shielding and other techniques to reduce the noise. It's not a new problem. An article from 1998 discusses it here. You can google for "emi cpu emissions" for more stuff on this.

Re:Again?

[Anonymous Coward]

>I do recall someone telling me that no CPU would ever run at more than 2GHz, as it would then start emitting microwave radiation...

Well... since the current generation of CPUs can already get hot enough so as to allow making some tea (or at least keeping it warm), I would not wonder if, in the near future, I could throw out the big microwave oven and instead use the CPU.

Re:Again?

[suv4x4]

I do recall someone telling me that no CPU would ever run at more than 2GHz, as it would then start emitting microwave radiation...

That's nothing, every time I drive my car faster than 20km/hour, I can't breathe...

Noore's Law

[Anonymous Coward]

The number of people predicting the end of Moore's law will double every two years.

And I predict that in 10-15 years time....

[Anonymous Coward]

We will start using a new technology without transistors, but which will exhibit a similar exponential gain so long as there is money to be made from it.

In fact, now I come to think of it, ALL human endeavour exhibits exponential growth so long as there is money to be made from it. Technology is just one field where it's true. Sex (Malthus), agriculture, you name it - humans do it exponentially!

I think I'll put that on my t-shirt, and call it 'Anonymous Coward's Law'.

Re:And I predict that in 10-15 years time....

[Colin Smith]

humans do it exponentially Life does it exponentially... Till it runs out of resource.

Re:And I predict that in 10-15 years time....

[Kelbear]

We require more vespene gas. There is not enough energy. Additional supply of depots required.

Re:And I predict that in 10-15 years time....

[Anonymous Coward]

And I predict that in 10-15 years time, there will be a poster on slashdot that can tell the difference between exponential growth and geometric growth.

Moore's law isn't really a law

[vlad_petric]

... there's nothing fundamental about it. Instead, it's a self-fulfilling prophecy. The big players in the silicon world all use the "law" and its corollaries as their business plan. They'll likely discard a feature/product if it falls behind the curve in terms of speed. For the layperson, this "precision" may indeed create the appearance of an actual law, even though it's just an observation (similar to Malthus' "law")

Re:Moore's law isn't really a law

[Anonymous Coward]

If I ate a nickel every time some Slashbot piped up about Moore's Law not really being a law, I'd be shitting $billions.

Re:Moore's law isn't really a law

[aadvancedGIR]

-As someone already posted, a law IS an observation (law of gravitation: objects attracts each other depending on their relative mass and their distance. theory of gravitation: something about gravitons, or something else, no one is really sure what it really is).
-Moore's law is just a particular case of learning law (any industry tend to regulary improve its production technique as long as the improvement has a positive ROI to justify investing money, wether the result is cheaper or better products is a choice left to marketing, not science).

I wonder

[packetmon]

Has Moore ever heard of Murphy?

Since When...

[His Shadow]

So many things in the tech world that should die (PS2, MBR) and we perennially have articles about a law that never was "dying". Cutting edge journalism, all right.

Re:Since When...

[Anonymous Coward]

My PS2 rocks! My PS3 on the other hand...

10 years ...

[polar red]

that means 2^5 ... that's 32 times more computing power on a single chip ... that's about 16 times the computing power I need.

Re:10 years ...

[mce]

Nobody will ever need more than twice the computing power he already has available on the moment he makes this assessment.

Re:10 years ...

[Anonymous Coward]

Why do you need twice the computing power but not more? Your current system is too slow for HiDef porn playback, but once that's fixed you have no other need?

Re:10 years ...

[kebes]

Well, it's about 1/1000 of what I would "need."

Of course, by "need" I mean what I would like to have. People keep talking about computers being "fast enough," but every time I have to wait for something to finish (whether it's a Photoshop filter, compiling code, running an optimization, or waiting for a Slashdot page with hundreds of comments to load), that's time I could have used a faster computer.

If my computer were 1000 times faster, then things that currently take minutes would finish "instantly." It's not just a matter of saving seconds here and there: having complex operations being effectively instantaneous changes one's workflow. (E.g. instead of trying different filter settings in Photoshop iteratively, it would be nice to have a live preview, even for the computationally-intensive filters; or, it would be nice to have complex multi-dimensional data-fitting finish so fast that you can try different models in realtime, rather than waiting hours for it to finish...)

All I'm saying is that there is indeed a use to having more and more computing power.

Re:10 years ...

[Atriqus]

Lets all come back to this post in 10-15 years so we can have a good laugh about it; Think about it, 15 years ago the 66 MHz Pentiums were first unveiled... Future Selves: "A Core 2 Quad was all he'll ever need?! Man that thing would barely be able to open a notepad nowadays; well, after it spent 2 months booting the operating system anyway. :)"

It's a law of econmics

[goombah99]

Moore's law is not about physics it's about economics. Basically the entire industry has built an economic engine that requires that growth pattern to sustain it self.

To put it another way, growth needs to be geometric not addative. that is things need to grow at x% per year, which leads to a doubling time. If the grew linearly at x += D then as x grew the proportional rate (1/x dx/dt) of x growing shrinks with time--or the doubling period gets longer and longer. Eventually it takes a lifetime before your computer is 2x more capable. Then it takes 2 lifetimes.

Why would you ever upgrade at that point? except due to wear and tear. Things become commodities and sales are based on price and other values-added. So long to intel's industry domination model.

Moore's law is also a limit too. Namely that very same growth engine will not invest twice as many research dollars to get a slightly faster doubling time. The fact that it has held steady tells you that this is so. Empirically this growth rate is the sweat spot between creating innovation at the lowest cost, and reaping a profit on it.

Indeed the only surprising thing we've seen in the consumer market that seemed (superficially) to violate this was apples' replacement of the ipod mini with the ipod nano shortly after it's introduction. They could easily have milked it for longer. But here the driver was the competition that they needed to stay ahead of.

Re:It's a law of econmics

[jcr]

Moore's law is not about physics it's about economics.

Exactly. Whenever one process technology reaches its physical limits, we get a new one, because the new process makes money. X-ray lithography, chip stacking, 3D circuits, and eventually nanotech will all keep us on the Moore's law path probably for the rest of my life, at least.

-jcr

Re:It's a law of econmics

[polar red]

so you're 75? 80?

Re:It's a law of econmics

[Jeff DeMaagd]

You are right, but that's also because the fabs get more expensive on each generation, I think each feature size shrink requires a fab that costs 50% more than the previous fab.

Re:It's a law of econmics

[goombah99]

You are right, but that's also because the fabs get more expensive on each generation, I think each feature size shrink requires a fab that costs 50% more than the previous fab. See my post below about the corollary for more discussion. But right now your point does not hold simply because the size of the market is increasing and revenues are also increasing. Therefore 1) cost-per-cpu cycle and the cost per unit of computation is falling despite the increasing cost of fabs 2) the growing cost of the fabs as a fraction of growing revenue is not increasing (I believe) yet.

Re:It's a law of econmics

[jcr]

No, I just feel that way once in a while. ;-)

-jcr

Re:It's a law of econmics

[Anonymous Coward]

It could also have something to do with chip designs being an NP-complete problem, and approximations taking polynomial time to compute. If the rate of performance increase was twice as much as it is, then we would be able to lay out and test new chip designs sooner leading to a faster rate of performance increase over what we have.

In short, the future rate of performance increase depends on the current rate. Regardless of process a 100 Mhz Pentium is just not usable for creating and optimizing the design of a Core 2 Duo, the design process would simple take too long. So even with somewhat major breakthroughs the Moore's law will accelerate or slow, not break.

Re:It's a law of econmics

[FuzzyDaddy]

Moore's law is not about physics it's about economics.

Your point on economics is well taken. However, there is one aspect of physics in Moore's law - that the equations governing a MOS transistor scale with size. That is, if you make a transistor that is 1/2 the size in all dimensions, and you run it at 1/2 the voltage, it will behave exactly the same as the original. So there has always been a clear development path for doubling your transistor density - cutting the size in half.

Other technologies (internal combustion engines, batteries, etc) improve in fits and starts - sometimes dramatically - but not by the continuous scaling of a single parameter.

(This is in no way minimizing the enormous technical challenges of making and designing smaller and smaller transistors. It's just that you know that if you can figure out the fabrication, you know it will work.)

Re:It's a law of econmics

[hackstraw]

Whenever one process technology reaches its physical limits, we get a new one, because the new process makes money.

I kinda agree and kinda disagree.

Moore's "Law" is clearly stated in terms of physics. It says that the number of transisters will double, not the speed will double over time.

However, as Kurzweil and other's have observed, the speed of _computation_ has doubled over time before Moore's law and there is no reason or hint that this will stop once Moore's law is obsolete.

Take a peek at http://www.kurzweilai.net/articles/art0134.html?printable=1 specifically http://www.kurzweilai.net/articles/images/chart03.jpg

ICs have been good for a while, but then so were abacus' at one time.

CPUs are simply different than they were a few years ago. Things like the Niagra chip from Sun and the multi-core stuff from AMD and Intel is pretty different design (SMP on a chip -- yes, that is an oversimplification).

10-15 years is about in the middle of 2020, which seems to be a common point of a number of interesting stuff. Physics computations are predicted to be pretty interesting by then. Computers are predicted to be interesting by then. Who knows what else.

Its not hardware that I think is the problem or challenge, its the pains of software that seems to be more challenging. I mean its 2007 and we have what for software? OSes and compilers and whatnot have pretty much stagnated since the early 70s. Sure, we have 4g languages that are easier for us stupid people to program with, but from a performance and efficiency POV they are backwards, not forwards. JIT stuff in .NET and Java are a little interesting, but programming computers is still a PITA.

I guess we will have to wait and see.

Re:It's a law of econmics

[Anonymous Coward]

Empirically this growth rate is the sweat spot between creating innovation at the lowest cost, and reaping a profit on it.

Sweat spots are for pubicly traded companies.

Re:It's a law of econmics

[KiwiCanuck]

Originally, Moore's rule was based on physics (now it is quite contorted). Here is the rule. The density &/or speed could double when the min feature size was reduced by 2. One of the reductions was decreasing the thickness of the oxide layer between the semiconductor channel and the gate electrode. This is where we hit a snag. When you are down to 3 atoms, you can't divide by two. This snag is also why the static power has rocket up over the past few generations. There is current leakage b/w the gate and the semiconducting channel is exponetially proportional to the barrier height. The separation difference b/w the gate electrode and the channel is a part of the equation. Another important part is the dielectric constant (the higher the better). Enter the high-K dielectrics. These are supposed to solve this problem. They create al physically larger barrier for an equivalent oxide thickness. However, the dielecetric constants are typically lower. So there is a trade off, the trick is to find a happy medium. If you are interested look up high-k dielectrics. The whole point is to increase the barrier.

Re:It's a law of econmics

[josephdrivein]

Of course, this is just the main concept.

There are a number of issues: e.g. scaling the supply voltage is expensive because it requires the replacing of the power supply, then there are a number of physical problems, e.g. subthreshold conduction.

See: http://en.wikipedia.org/wiki/MOSFET#Difficulties_arising_due_to_MOSFET_scaling for more.

Re:It's a law of econmics

[Miamicanes]

When the limit of Moore's law is reached insofar as doubling of computing power for a given size, things built from the components in question will simply start getting bigger and drawing more power. The Pentium 4 was just a preview of what's to come. Forget about tech-design fantasies of "computers in a keyfob" that go everywhere. The desktop PC of 2050 will occupy a case that would have given a 1995 power user a serious case of "tower envy", draw more power than the half-dozen halogen floodlights illuminating the back yard to stadium levels, and an active cooling system with outdoor condenser that needs to be connected by refrigerant-containing hoses through a hole in the wall or a partly opened door/window. Of course, it'll have the equivalent of 65,536 discrete CPUs, a terrabyte of onboard ram, up to a petabyte in nonvolatile storage, a GPU subsystem that's more powerful (and draws more power) than the CPU array itself (needed for immersive 3D games if you want to use a haptic bodysuit & gyroscopic G-Force simulator platform along with it).

Believe it. The industry won't come to a halt when exponential growth of component power ends... it'll just embrace the exponential growth of size and power consumption, and convince consumers that they simply MUST have a computer the size of a small refrigerator that renders apps into solid-feeling virtual tablets (haptic glove & visor required for full Genuine Windows Experience) that can float, spin, and (if you angrily hurl one at a wall) smash into a million pieces before turning into pools of mercury and slurping back into the computer in a cute "shatter" effect.

Re:It's a law of econmics

[Anonymous Coward]

Empirically this growth rate is the sweat spot between creating innovation at the lowest cost, and reaping a profit on it.

Is the sweat spot anything like the wet spot? If so, it's no wonder Moore's trying to avoid it!!1 Nobody likes to sleep in the wet spot.

You probably meant the sweet spot.

Re:It's a law of econmics

[GWBasic]

However, as Kurzweil and other's have observed, the speed of _computation_ has doubled over time before Moore's law and there is no reason or hint that this will stop once Moore's law is obsolete.

By the time Moore's law runs out for Silicon, we will see it sustain for non-silicon-based computing.

Re:It's a law of econmics

[joto]

But right now your point does not hold simply because the size of the market is increasing and revenues are also increasing.

As long as you keep putting out better electronics components at a lower price, the market will keep increasing. As Moore's law continues to do its work, new products will be made possible, and because the people who makes new gadgets first tend to profit, the market will ensure that new products are made. Here is a chronological list of some of these gadgets that almost overnight went from unthinkable (or at least hugely expensive) to normal household items: the calculator, the PC, the gaming console, the luggable PC, the CD-player, the pocket gaming console, the mobile phone, the DVD-player, the laptop, the digital camera, the mp3-player, the GPS.

20 years ago, people didn't even see the "car entertainment system" as having a place in science fiction. Today you find them in the bargain bin in the supermarket. And it exists purely because of Moores law.

Best part about predicting own failure

[gurps_npc]

Is that even if you are wrong, you are still right.

Wow, that Moore guy was so smart he outsmarted Moore.

Hafnium Breakthrough

[smitty97]

from tfa:

"We're not far from that," Moore said on Monday. "Before we had our Hafnium breakthrough, we were down to the point where we were at five molecular layers in the gate structure of the transistors. When you clearly can't go below one...you get into other types of problems," Moore said. The law will continue, they just need breakthroughs with Quarternium, Eigthnium, etc..

And next year ...

[mshmgi]

Next year, they'll tell us that Moore's Law will end in 5-7.5 years.

Re:And next year ...

[ttapper04]

I thought your comment was very funny. I would mod it if I had the points.

CPU speed already on the wane as consumer bait

[dave_mcmillen]

I have no idea if Moore's Law will really start to "fail" in a particular time scale (one of these times it's gotta be true, right?), but a related issue I find interesting is that CPU speeds don't seem to be being touted to computer buyers so heavily anymore. Walk into a big electronics store and look at their desktop offerings: where they used to prominently feature how many GHz they had inside (and people vaguely felt that more of these mysterious GHz was better), now the CPUs are given code names and numbers that don't reflect CPU speed: Check out this nifty X2, or the Turion 64, or ...

The new hook for consumers is the number of "cores", and once again most people have probably picked up the vague sense that having more of them inside means the computer is better. I've been told by people who might be in a position to know that it's not that they can't keep cranking up CPU speeds, but that the cost/benefit (profit-wise) stops making sense at some point because of the huge cost of implementing a new fab at a finer length scale, and we're pretty much at that point. So it makes sense that cores are the new GHz, and Moore's Law will have less and less direct impact on the end computer buyer from now on.

Maybe there's a Core Law to be formulated about how often the average number of processors per computer can be expected to double?

Re:CPU speed already on the wane as consumer bait

[Cutie Pi]

As the parent implied, Moore's Law will likely not end because of technological constraints but rather economic ones.

We reached a wall a few years ago in terms of transistor speed, mostly due to the thin gate oxides giving rise to significant leakage current, which translates into heat. The upcoming high-k metal gate technology mitigates but doesn't eliminate this problem. Thus, Intel and the like are putting those smaller transistors to work in redudant cores rather than faster, monolithic circuits. However, Moore's law is still marching on, from 65nm, to very soon 45nm, and then 32nm and 22nm. Each technology node effectively halves the area requirement, or (more realistically) doubles the number of transistors that can fit in the same area. This translates into lower cost per transistor. 32nm technology is in the final stages of development, and 22nm is believed to be possible, although much more difficult. The real limit right now is the optical lithography process used to pattern the circuits. There is no high-volume solution available for the 16nm node. We can certainly make patterns this small (using electron beam lithography, for example), but it would be prohibitively expensive (each chip would take many hours to expose, compared to SFIL but the industry isn't really putting money into them. My belief is that the 2-year cycle we're on will be ending after 22nm, or possibly 16nm. Circuits will probably continue to get smaller, but at a slower pace, as development and technology costs become prohibitive. Even now, there is debate in the industry about whether to delay 22nm and instead do a 28nm "half node". If this was to occur in the entire industry, the 2-year Moore's Law as we know it would end.

BTW, "45nm" and "32nm" don't directly refer to the size of the transistors, but rather to 70% of the current half-pitch being printing with lithography. Thus, the 45nm node has a 65nm half-pitch, which means the wires are spaced at 130nm. This spacing decreases to 90nm and 65nm for the 32 and 22nm nodes, respectively. The actual transistor size (channel length) can be much smaller than the currently technology node designation.

Re:CPU speed already on the wane as consumer bait

[TheRaven64]

CPU speed isn't necessarily a draw, but power consumption is. Power consumption is something most computer buyers care a huge amount. Mind you, most computer buyers call their computer a 'phone' these days for some strange historical reason.

Re:CPU speed already on the wane as consumer bait

[cowscows]

Moore's Law doesn't really have anything to do with MHz or GHz, or clock speeds at all. It's more about the number of transistors crammed into a cost effective chip. For a while there, one of the main things that intel decided to do with all those transistors was to push the clock-speed as fast as they could. This certainly made computers more powerful, and it was an easy number to work into advertisements and such.But at the end of the day, it wasn't the only way that processors could be improved. To bring in a dreaded car analogy, they were making a car go faster by adding more cylinders to the engine, while mostly ignoring things like aerodynamics and efficiency in other parts of the car. But eventually they reached a point where there wasn't any room in the engine compartment for more cylinders, so now they're looking at making the rest of the vehicle more efficient.

The transistor count will keep going up, Moore's Law will continue. It's just that those new transistors will be used a little differently.

Re:CPU speed already on the wane as consumer bait

[Frenchy_2001]

If you think that Moore's Law is about the frequency of the processor, you are badly mistaken (but it's mostly not your fault as this is what it has been summarized to in the media).

Moore's Law is a law of economics, scale and progress.
The gist is that computing power at a given cost will double every 18 months. It does not matter if this progression is achieved by cranking the frequency (MHz rule) or by increasing the number of transistor and parallel processing (Core rule). This is all about the economics of processing power.

It comes from the fact that the industry transitions to a new, finer technology every 12-18 months and hence can build more, faster transistors in a given area of silicon.

The problem is that each new transition comes with new challenges and as long as the challenge is not overcome or at least has a roadmap to a solution, there is a risk the progression stops here. This happens every few years. This mostly tells us they have plans for progressions for the next 10-15 years and unless we have newer discoveries in the mean time (which has always happened so far), we will not overcome this limit. 10 years is a long time in term of scientific discoveries...

Re:CPU speed already on the wane as consumer bait

[jafuser]

The new hook for consumers is the number of "cores"

Pfft, you early cent gramps and your "cores". Just pend for 2k50 when you can viz into the main on your very own petaqux rez. =)

NO NO NO NO NO!!!!!

[Colin Smith]

Moore's law will continue until THE SINGULARITY takes US ALL!!!!!!

Or at least, that's what the singularity nuts claim. Sorry people, there are limits on this planet.

Re:NO NO NO NO NO!!!!!

[arevos]

The singularity nuts essentially claim that, (a) it's possible to build a better-than-human AI, and (b) it's possible to get a lot more intelligent than humans currently are. Neither assertion seems particularly far-fetched to me.

Re:NO NO NO NO NO!!!!!

[Colin Smith]

The singularity assumes intelligence and knowledge can increase basically infinitely. Hence, singularity... Which is utter bollocks.

It may well be possible to make a machine more intelligent than a human... God knows most of us are stupid, but that does not make a singularity.

Re:NO NO NO NO NO!!!!!

[DragonWriter]

The singularity nuts essentially claim that, (a) it's possible to build a better-than-human AI, and (b) it's possible to get a lot more intelligent than humans currently are. Neither assertion seems particularly far-fetched to me.

Actually, I don't think that at least the "key" singularity "nuts" (Kurzweil, etc.) necessarily claim that, except using "AI" in a very loose sense. As I recall, Kursweil's big claim is essentially that human intelligence + computer power is effectively a greater form of intelligence, and that the explosion of computing power itself feeds the creation of all kinds of new technology, including an even faster explosion of available computing power, which feeds back on itself with an increasing proportion of the information processing necessary to synthesize new knowledge from existing knowledge being done outside of the human brain. IIRC, Kurzweil is a strong AI proponent as well, but that's largely tangential to the singularity issue, which is mostly dependent on the positive-feedback loop in computing technology and the ability to record and process information to synthesize new knowledge. The issue is more the multiplier effect that computing technology provides to human intelligence than either "AI" in the usual sense or increasing innate human intelligence.

I could be misrembering, though.

Re:NO NO NO NO NO!!!!!

[gravos]

No. The "singularity" is caused by an increase in intelligence beyond human levels that obscures our ability to make meaningful predictions. There is no need for any kind of infinite growth. Please do not discuss things you do not understand.

Re:NO NO NO NO NO!!!!!

[Anonymous Coward]

"Please do not discuss things you do not understand." This is Slashdot. Discussing things not understood is what happens here.

A law is an observation

[benhocking]

To simplify things a bit, a law is an observation, whereas a theory is an explanation. They are not the same thing, but you can have laws and theories dealing with the same subject matter.

Re:A law is an observation

[fbjon]

To use a bad analogy: A law is like a traffic law you have to follow strictly. A theory explains who enforces it.

Details at 11

[IorDMUX]

In other breaking news, the Department of Homeland Security reminds everyone that the terrorism thread advisory level has been raised to 'orange'.

Moore's Law will expire

[CruddyBuddy]

The end of Moore's Law is 10-15 years away

just like "true" AI.

cost per transistor

[Anonymous Coward]

Transistor density may reach physical limitations after we shrink everything to atomic scales and construct 3-D wafer-bonded chips, but the original statement of Moore's law was about the cost per transistor. It is likely that long after we stop increasing transistor density, breakthroughs in fabrication processes such as molecular self-assembly will lead to a second-wind for the exponentially decreasing cost per transistor. The computing industry will experience exponential improvements in throughput-per-dollar long beyond the end of semiconductor scaling.

Depends...

[john_is_war]

The true death of Moore's law will occur when we mastered quantum transistors. Once we hit that level of technology, I highly doubt we can progress anymore. Or at least as the same pace.

Re:Depends...

[Anonymous Coward]

You really have no imagination.

two variables

[OwlofCreamCheese]

he forgets his law has two variables, the number of transistors and the cost. The number of transistors might stop doubling, but there is still huge change to come with the transistors being fixed and the cost changing.

I mean, a 8 core chip would be an improvement right now, but so would a 4 core chip at half the price. Think about a world where a 80 core chip exists, and how it would change the world. It would do so again when it went from 999 dollars to 99 dollars to 9 dollars to 99 cents to 9 cents to 9 for a cent/

Re:two variables

[co1dfus1on]

10 for 9 cents not 9 for a cent (9/10 != 1/9). (It would actually go from $1024 to $512 to $256 to $128 to $64 to $32 to $16 to $8 to $4 to $2 to $1 to $0.50 to $0.25 etc..).

Re:two variables

[Rodyland]

Hard to think about how an 80-core chip will change the world when the second core on my desktop spends most of its time idle.

But I'm sure as we move from 2, 4, 8 an upwards core chips the software will evolve to make use of it, but I would suggest that whatever it is you'll be doing on your 64-core chip in 10 years time hasn't even been invented yet.

I personally think it'll be the interplay and interactions between the big (64-core multi-GHz) and the small (phone, pda, ipod, dvd, tv etc) that'll make things interesting.

Moore did not come up with exponential growth

[Anonymous Coward]

The uplink poster seem to imply that warios exponential growth laws (Malthus, Compound Interest, etc) are particular cases of Moore law. WRONG! Exponential growth laws have been known for centuries, from the Middle Ages, even earlier. The law of compound interest was well known to Italian merchants in the Middle Ages (remember Fibonacci!). Malthus did not write any equations, the exponential law for population growth was introduced by Euler half a century before Malthus. Moore law is no big deal, just a late example of exponential growth.

Re:Moore did not come up with exponential growth

[Anonymous Coward]

I like to attribute the growth to Mario, but I guess Wario played a role too.

Not yet,

[SharpFang]

The law speaks about number of transistors. Considering current size of a typical CPU die (about 1cmx1cmx1mm) and assuming a "reasonable" maximum size of some 10cmx10cmx10cm we have about 15 years of doubling the SIZE of the CPU (with some challenges like heat dissipation, but nothing nearly as difficult as increasing the density further) and that's not considering increasing the density any more. So even if the density reaches its limits, the CPUs may simply grow in size for a good while.

Re:Not yet,

[mce]

Sorry, but 10cmx10cmx10cm is not reasonable. Not only because you run into functional yield issues, but also because you can only fit so many dies of that size and form factor on a single wafer. And it ain't many. Manufacturing cost would skyrocket even in case the yield would be 100%, which it never will be.

Re:Not yet,

[tomstdenis]

A typical processor is not 100mm^2, they're more like 140mm^2 (core2duo is about 143mm^2 according to a quick google search). Do you have any idea how expensive a 1000mm^2 die would be? That would cut yields down, and also be really hard to package, not to mention that if you packed 1000mm^2 with 65nm transistors the thing would consume 300 Amps.

Sure I suppose it's possible to make a 1000mm^2 die, but it would cost $25,000 USD and probably either be a super-many-core or run really slow (think longest wire...).

Re:Not yet,

[Chirs]

You've neglected to consider power issues. The 10cm cube you mention is 10000x the volume of the "typical" current size you mention. Assuming power scales linearly with volume, that would require approximately 300KW of power just for the cpu. That works out to about 1250Amps of current at 240V.

Nobody wants to increase the size of cpus...defects scale more than linearly with area, so there is a strong incentive to keep the die area down. Also, as the physical size increases you run into other problems...power transfer, clock pulse transfer, etc.

Re:Not yet,

[TheRaven64]

Sure I suppose it's possible to make a 1000mm^2 die, but it would cost $25,000 USD Of course, according to Moore's Law, in 18 months it would cost $12,500. That's rather the point. Even if we stop being able to shrink dies, the process technology will become better understood and cheaper, and so we will be able to make more of them for a fixed investment.

Re:Not yet,

[jambox]

10cm x 10cm? No chance. Not in a million years.

One of the major considerations of designing a CPU is skew - the time it takes for a pulse of electricity to get from one bit of the CPU to another and back again. If the two bits are too far apart, you have to wait for the signal to return before you can carry on.

You can make a chip that big but maximum clock speed would be very low.

I can't believe that this parent got modded +4 insightful - anyone who has done even a basic architecture course would know that clock speed is inversely proportional to the chip size.

Re:Not yet,

[SharpFang]

If you consider 1 core being the size of 10cm, sure.
But I assure you a nowadays cluster of 1024 CPUs has the cores way further than 10cm apart from each other. Sure the computational power doesn't scale up in direct proportion to the number of cores, but still the performance gain is significant.

Re:Not yet,

[SharpFang]

Sure these are problems to overcome.
1) Power. We are right at the beginning of the road of scaling power requirements down. So far the power per CPU kept going only up. It's only the last two or three years where we're counting MIPS per Watt, not just per dollar. There's still a long way for the power to go down. There's a lot of ways to aid power dissipation without actually wasting the power - imagine integrating CPUs with utility water heaters for example, and as computers become the commodity more and more, this isn't entirely as silly as it might sound.
2) Defects are not an accident, they are a business factor in nowadays computers. You get x% yield off a wafer, you get faster or slower CPUs from better or worse series of the same run, and in tests you discard the failing parts. Now imagine you're not getting a solid brick of, say, 4096 fully operational perfect cores. You get a stack of wafers, each with between 40% and 98% (depending on price) of the cores working. The broken cores are cut off by quite simple and fault-proof part of the electronics. Something very similar is going on with gfx cards nowadays. A "LE" version of a card has exactly the same core as a "GT" version, except it has some of its shader units disabled, because they are faulty.

As for the other problems - sure we will run into them. Doesn't mean we can't get through them. The ones you've listed aren't nearly close to showstoppers, merely obstacles.

Re:Not yet,

[SharpFang]

Sure computers with more than a hundred million logical gates are impossible because they would require over a billion of transistors (each about half a centimeter big) and require half the Sun's energy output to run... Therefore I predict the increase in computational power of the computers will stop around year 1970.

A 10,000mm^2 die would cost as much as 100 100mm (10x10) dies, and would have as many faulty cores as you get from a standard wafer. If I packed 10,000mm^2 with 65nm transistors and cut the frequency to 1/8, the thing would consume about 50 Amps (power is proportional to speed squared). Still giving some 10 times the CPU power of a single core, and scaling the power consumption down very gracefully.

Now start stacking these on top of each other, with a coolant gaps between them.

Re:Not yet,

[SharpFang]

The biggest silicon wafers are some 300mm. This would fit 4 units of 8x8cm, a core 1cm big, that is 4 arrays of 256 cores each.
Stack 64 layers on top of each other. That makes it 16 wafers to produce one "brick" of 4096 cores.
Manufacturing costs would scale up directly with CPU power delivered and down with technological progress and investment returns.
Yield would be 100%. Efficiency of a single "brick" and its price would be a factor of the number of cores that are faulty and disabled in firmware - not all of the 4096 cores in the example would work. Expensive models would have over 4000 cores active. Cheap models would have less than half of them working.

Re:Not yet,

[mce]

Sorry, but 4 units out of a 300mm wafer leaves a big waste "around the edges". Of course you can try to fill that with smaller chips, but then having a non-uniform wafer adds complexity. So you're better off with 4x4cm units, of which you can put several extra on a single wafer. Now repeat the same logic until you reach the economically optimal point. I don't know exactly where that is, but is sure is not at 8x8, IMHO. Also don't forget that certain parts of the production costs are essentially determined by the amount of wafers, not the amount of dies. The less dies you get, the larger the part of those costs each of them has to carry. And since the semiconductor industry is ruled by very low margins...

A yield 100%? Have you ever been near / inside a fab? I know I have. I used to work in a company that has two of it's own, including one of the first 300mm ones. For starters, you're conveniently redefining yield to mean "as long as part of the circuit works, the die works". That's not correct: you may be able to sell a partially working one, but you also need a sufficient number of fully working ones to meet demand for those markets, otherwise the whole price segmentation plan breaks down. Next, depending on the exact fault, the entire die may be out of order due do a simple fault in a critical place. now, before you say "yes but redundancy...", please be aware that the bigger the die, the bigger the probability that it has multiple faults. In short, there simply is no such thing as 100% yield.

Re:Not yet,

[SharpFang]

...which are all obstacles and problems but not showstoppers.

Actually, I don't see any big problem using smaller wafers (say 8cm diameter) and filling them whole with the circuits in a circular pattern (nobody says a single core needs to be square!). Then make the CPU to be a cyllinder instead of a cube. No waste on the edges.
The yield will never be real 100% but redefining the meaning of yield into "percentage of non-flawed parts of a single die" as opposed to "percentage of dies with no flaws at all" would be a significant boon for the industry. And given 'sufficient performance' you don't need '100% performance'. A demanding buyer may want a gfx card with all 16 out of its 16 vertex shader units working. But once you get 512 shader units on a single die, you may sell the "GT" version of a card defining it as "more than 500 shader units", "standard" as "with over 400", and "LE" with "Between 250 and 400 units working". You may even adjust the price to performance in semi-linear fashion, just like with CPU speeds.

Of course the approach "1 die = 1 wafer, 1 CPU = 100 dies", is far more expensive than current technology, but doesn't mean with technology getting cheaper it won't become affordable. Build assembly lines which produce assembly lines of CPUs, silicon is plentiful, you just need lots of power to purify it. The density of circuits is a barrier we are reaching rapidly, but the costs of the dies in current technology can still be pushed a long, long way down.

Compare costs of simple electronic gadgets like a hand watch from 20 years ago and now. The technology is still the same but you get the watches in cereal boxes. Technology goes ahead, price stays the same. Technology stays the same, price goes down, meaning you can simply load more pieces of the same technology into a larger volume for the same price. You don't need separate 1000 hand watches which currently cost what one did back then, but just imagine a beowulf cluster of these ;)

My prediction

[Random832]

I predict the number of predictions of the end of Moore's Law will double every six months.

Obligatory Pedantry -- it's about what's cheap

[Urban Garlic]

Luckily, there are enough geeky pedants on slashdot to make up for the fact that the editors have actually messed up this totemic bit of geek lore.

Moore was/is a technology manager, and his law is a management law. It says the number of transistors that can be economically placed on an integrated circuit, i.e. the transistor density of the price/performance "sweet spot", will increase exponentially, doubling roughly every two years.

The original refers to "complexity for minimum component cost", which emphasizes the economic aspect of it even more strongly.

Moore's law has never been about what's possible, it's always been about what's cheap.

Re:Obligatory Pedantry -- it's about what's cheap

[owlstead]

Just because it doubles because of economics (which I highly doubt), that still does not mean that it cannot run into technical difficulties. As a previous poster said: it's not like the fuel economy of cars follows anything like Moore's law, not even a seriously slowed down version of it. And you won't be able to run a car on just a few Joules at the end. In other words: there's economics, and there is the real world.

Corollary to moores law

[goombah99]

If you accept the statement I just made about moore's law being sustained because of economics then here's a corollary which makes an observable prediction.
Moores law stays fixed because the industry invests enough research dollars--and not one dollar more-- to keep it at that rate. Their entire economic model is built on this.

Therefore, if we every do reach a point where we simply are running out of available physics and computer science (multiprocessing) then the first sign of this will be an increasing fraction of research dollars spent to sustain moores law.

Plot the industry's margin, smooth the curve, and you will be able to extrapolate to the point where the research dollars cross the profit line. somewhere shortly before that is when moore's law will end.

The only way that would not be true is if the nature of innovation changes from frequent small leaps to massive leaps spaced far apart.

Re:Corollary to moores law

[Bob-taro]

Moores law stays fixed because the industry invests enough research dollars--and not one dollar more-- to keep it at that rate. Their entire economic model is built on this. What makes you say that? What about competition? If you knew "the other guys" were striving to exactly meet Moore's law, wouldn't you try to beat it?

Re:Corollary to moores law

[Henry V .009]

Plot the industry's margin, smooth the curve, and you will be able to extrapolate to the point where the research dollars cross the profit line. somewhere shortly before that is when moore's law will end.

Because the future can always be predicted from current trends...

Re:Corollary to moores law

[MarsDefenseMinister]

39 years ago, I didn't exist. currently there is one example of me existing. Therefore, 39 years from now there will be two of me!

Re:Corollary to moores law

[goombah99]

Moores law stays fixed because the industry invests enough research dollars--and not one dollar more-- to keep it at that rate. Their entire economic model is built on this. What makes you say that? What about competition? If you knew "the other guys" were striving to exactly meet Moore's law, wouldn't you try to beat it? No it's called a Nash Equilibrium. A point in competition space where no player can imporve his strategy given the other players moves. I can't say what the costs that drive it are. But it's so fixed it apparently has reached equilibrium.

Re:Corollary to moores law

[Bob-taro]

No it's called a Nash Equilibrium. A point in competition space where no player can imporve his strategy given the other players moves. I can't say what the costs that drive it are. But it's so fixed it apparently has reached equilibrium.

So are you saying it's a Nash Equilibrium because the technology has followed Moore's law so closely that therefore all the companies must be doing just what they need to make it true and no more? I'm not arguing, just honestly trying to understand your point.

Re:Corollary to moores law

[crgrace]

That is a good statement, but unfortunately people have been doing exactly what you suggest since at least the late 1970s. The problem is, there is no good way to predict or even fathom what innovations will drastically change whatever assumptions have to be made to enable the forecast in the first place.

There were grave technological barriers to making transistors with gate lengths smaller than 1 micron. It was not clear in the early 80s if they could be solved. In fact, they have been solved and at a cost much, much, lower than really anyone predicted. This has happened several times in the semiconductor business.

Now, the semiconductor roadmap is assuming that some technological innovation will enable us to make deep UV lithography an econonmic reality. It is behind schedule, and direct e-beam lithography is far to expensive to use to pattern wafers (it is used in mask fabrication at present). Maybe the end is nigh. Based on history, I think there will be a work around. There is so much money at stake.

This whole Moore's law obsession is not germane. This is a business. When it makes sense to continue to pour billions into semiconductor manufacturing R&D we will. When it doesn't, we will stop. It's quite simple. It is common knowledge that the improvements in software have mainly come from faster computers (Boyle's law of software and all that). When it is too expensive to continue to improve semiconductors, we will do something else, and it will become cost effective to do more research into software.

Re:Corollary to moores law

[goombah99]

Yes that's my argument. The main argument against it is that there simply are not enough companies trying enough different strategies to have fully explored the competition landscape. Could be flawed. But I think it's on target. (and don't call it a circular argument unless you have a better explanation of how despite the fickle nature of experimental prgoress moores law has sustained it self well over 9 orders of magnitude.)

Nash Equilibrium in a nutshell is something every efficient industry will always reach. It's game theory if you a familiar with that.

Re:Corollary to moores law

[joto]

Moores law stays fixed because the industry invests enough research dollars--and not one dollar more-- to keep it at that rate. Their entire economic model is built on this.

What makes you say that? What about competition? If you knew "the other guys" were striving to exactly meet Moore's law, wouldn't you try to beat it?

Beating Moores law can only be done by the semiconductor companies, not the companies building end-user products from chips other companies creates. These companies know about Moores law, and start planning their products so that they know that at least one of the semiconductor companies will have a product that fits their need in e.g. 16 months. While having a chip available 3 months earlier is nice, it doesn't really matter much, since it (a) screws up planning, and (b) most likely is too expensive to use in a consumer product. Since the IT/technology-sector is often a relatively long chain of suppliers and buyers, selling to each other, untill you finally reach the end-user, Moores law is the yardstick that is used for all planning that needs to be done.

Ob Simpsons quote

[Maximum Prophet]

Homer: Lisa, in this house, we obey the laws of thermodynamics!!!

Shriver's Law

[jshriverWVU]

Computing power will continue grow in direct relation to finite amount of knowledge we have regarding physics. For each advancement in our knowledge of particle physics the more apt we are to apply it toward electronics in general.

Duke Nukem Forever...

[Notquitecajun]

Hmmm...that coincides with the release date of Duke Nukem. Give or take 3-20 years.

Re:Duke Nukem Forever...

[El_Oscuro]

But will it run on Linux?

Nope, nope, and nope

[Ancient_Hacker]

First of all you've misquoted Moore's law.

Secondly it's not so much a "law", as a consequence of how long it takes to amortize the cost of a fab plant.

Thirdly, it's tied to 2-D circuit layouts. If and when 3-D IC technology becomes practical, then all we need is 2^1/3 percent or about 22% linear shrink every year, which is somewhat more maintainable for a few more generations.

10 years is 5 more cycles

[hattig]

That's 32 times as many transistors... whereas today you can get 4 cores on a CPU in under 300mm^2, you'll be getting 128 cores in 2017 (simplistic, you'll get a variety of generic cores, and application specific cores, and per-core improvements will increase their size, so say 32 generic cores and 32 application specific cores).

If it's 16 years, thats 256 times as many transistors. 256 generic cores and 256 application specific cores in 2024? Let's not even imagine the per-core speeds! It's all pretty exciting, and I'm being conservative with the figures here.

Of course, applications will grow to utilise this stuff, but more and more tasks are getting to the point of 'fast enough', even despite the bloating efforts of their creators. Even if there is a 10 year hiatus in process improvements after 2024, it'll take some time for the applications to catch up apart from certain uses. If those uses are common enough, there will be hardware available for it instead. Of course if only Intel and IBM have fabs that can make these products, because the fabs cost $20b each...

Re:10 years is 5 more cycles

[Anonymous Coward]

Of course, applications will grow to utilise this stuff, but more and more tasks are getting to the point of 'fast enough', even despite the bloating efforts of their creators.

I'm going to call your bull here. Applications will grow, and some will scale to use more cores. Some applications are "embarrassingly parallel". Most are not. Optimizing compilers might extract enough threads to fill 4 cores, maybe splitting an application into obvious "components" will pull a bit more, but all of this is expensive to do. Ever try writing parallel code that does something worthwhile, has good scaling, and is correct? Most experts can't even do that.

The computer industry is a speeding car headed for a brick wall right now. We don't know how to improve performance of current applications, and we don't know how to make applications that can scale in the future. Research is being done, but it takes a long time for research to trickle down into real life. Within a few years we'll be using 16 core chips that may utilize 4 cores...

Re:10 years is 5 more cycles

[Paul Fernhout]

The hopeful social outcome of all this increase in productivity was talked about as far back as 1964:
http://www.educationanddemocracy.org/FSCfiles/C_CC2a_TripleRevolution.htm
in a letter sent to President Lyndon B. Johnson in March 1964 called "The Triple Revolution".

Actually, the increase is more like a doubling every 1.5 years, which is about seven cycles in ten years, or more like 128X. But the rate of increase itself has been increasing too. Price has also been dropping. This makes effectively a 1000X increase in price/performance per decade at the current rates.

By the time any toddler of today is finishing graduate school, computers will be about 1000X (for the first decade) multiplied (not added) by 1000X (for the second decade) or about a million times faster than they are now -- just like computers are about a million times faster than twenty to thirty years ago (at constant dollars, or so MIPS per $). Related links:
http://en.wikipedia.org/wiki/Moore's_law
http://www.kurzweilai.net/articles/art0134.html?printable=1
http://www.bootstrap.org/dkr/discussion/0126.html
http://www.transhumanist.com/volume1/moravec.htm
(The rate of exponential growth itself is even increasing!) According to that last link, those AI computers had about 1 MIPS processing power. (And it's a funny idea Hans Moravec had, and I think correct, that only for the last decade or so has AI been taking advantage of faster desktop CPUs going beyond 1 MIPS..)

At lower previous rates, over 30 years, we see a million times improvement. As an example, compare the late 1970s Apple II
http://en.wikipedia.org/wiki/Apple_II
with todays' (2007) eight core Mac Pro.
http://www.apple.com/macpro/
Then --> Now (approximate increase)
CPU: 1 Mhz --> 8 * 3 Ghz (8000X faster, but about another 100X internal improvements from wider data operations and pipelining and such). (somewhere in x100000 to x1000000)
RAM: 4K --> 4GB RAM just starting to be common. (x1000000)
Disk: 300K disks --> 300 gigabyte disks. (x1000000)
And all for about the same price (adjusted for inflation). Some other considerations:
Bandwidth: 11 bytes/sec modem at $10 / hour --> 800000 bytes/second by cable at $60 / month (about x10000 faster, well that doesn't quite fit, but its still a big improvement -- and if you factor in the cost for continuous access, there is probably another 10x or 100X boost in there, producing effectively close to a x1000000 improvement of price/performance)
Printing: about 1000 characters per minute for $1200 printer -> 10 pages per minute each with millions of color pixels -- with the printer often now free with the computer (not sure how to call this as a multiple, since quality has changed so much).

So, here are possible specs for a personal computer of 2027 if it was a million times faster than today's:
CPU: 8 * 3 Ghz --> 8000 X 3 THz (1000X more CPUs each 1000X faster, though I think it likely such systems might just instead have a million processors at about today's speeds, perhaps interweaving memory and processing power)
RAM: 4GB --> 4000TB (enough to hold all of the current surface internet in RAM, see:
http://www2.sims.berkeley.edu/research/projects/how-much-info-2003/internet.htm )
See also:

Then it is..

[Anonymous Coward]

.. the "prediction of gravity"?

It's not a law it's an observation

[gelfling]

It's not a law, it's simply an observation that within Intel, that's more or less the rate of progress. As we saw with the P-4 chip the problem we bumped into was not Moore's Law, but the laws of thermodynamics. So we found a good enough reason to go to multicore CPU's. Eventually though you do bump into Albert Einstein. In 1 billionth of a second, light travels about 1 foot so the entire circuit length from end to end, in order to have a switching frequency of 1 billionth of a second, has to be less than one foot.

I Fought Moore's Law And The Law Won

[tiktok]

Why buy a computer this year, when I can get a faster one next year?

The fuzzy logic behind not buying a computer due to Moore's Law.

We're already there

[Animats]

We're already near the end of Moore's Law. The problem is not feature size, it's getting rid of the heat. CPUs are already hitting heat and power limits, which is why CPU speeds stalled out around 3GHz.

Feature size alone matters for memory devices, and we can expect continued progress in memory density. Even for DRAM, getting rid of the heat is becoming a problem, so the future is with devices that don't require refresh cycles. We'll see progress in flash memories and static memory technologies.

Moore's meta-law

[Weaselmancer]

The number of predictions about the end of Moore's law will double every two years.

Re:Moore's meta-law

[Anonymous Coward]

The number of jokes about the number of predictions about the end of Moore's law will double every two years.

Python

[goombah99]

IN ten years, according to moore's law python will be 32 times faster than it is now. Right now it's about 1000x slower than tuned C and 100x slower than unoptimized C-code. So in ten years python will still be slower than C running on todays computers. (mean while C will also be 32 times faster).

That is not a rag on python. well no a big one. indeed for lots of things don't need to be faster (word processors) so being 32x faster would enable python to take over development of lots of areas we now use C for.

No the point here is that if python can't even see a future where it's faster than today's other languages before it's obsolete it needs to go on a vision quest. Other than taking over distasteful roles that C refuses to do anymore, what's the point in life?

I think, like the thinking-in-python dude's rant said--python needs to ask it self what high level languages could be really good at that low level laguages will perpetually suck at. And that is multi-processing. thread safety is possible in any language but if you actually are thinking about while you are programming then you have a problem. Too hard. If you modified C to be thread safe intrinsically it would dramatically slow down. But if you modified an already slow language for this then it's not going to make a big difference in speed. Thus proportionally high level language poise to gain the most advantage by multi-processing.

And moores law is going to vector in to multi-processing in the future as a way to sustain itself.

Python should reinvent itself to be the multi-processing language.

Otherwise things like Fortress, which everyone scoffs at these days, is going to go to charles atlas school and be kicking sand in all your faces. (fortress is written from the ground up to assume multi-processing by default: e.g. for-loops always can execute in any order and the local variables are thread safe.)

Re:Python

[Anonymous Coward]

man, why is it that crazy people always feel the need to just fucking ramble on for paragraph after paragraph?

Re:Python

[Surt]

You're assuming that in 10 years they'll have made no optimizations to python? I mean, maybe, I don't much like that language and think it is probably a dead end, but even the currently less popular ruby is getting more than twice as fast at basically everything in the next version. Surely python programmers can get it at least 50% faster?

Re:Python

[goombah99]

I exaggerate slightly for the sake of being funny. But not much. Python is trying to optimize itself by adding more and more features or libraries for more very special cases that allow speed for that case. I think that has limited potential. And worse I fear adding features which need be supported in future versions may slowly foreclose the right direction which is to 1) permit implicit and explicit typing so that it can become a compiled language 2) create ways to do multi-processing.
So yes python can get faster but not much I expect.

offtopic:
personally I'm eyeballing Groovy right now because it is a script language that mirrors a compiled typed language that already has some degree of thread safety and behind the scenes memory management (GC) at a low level: java. But I'm still using python (and perl!)

Re:Python

[joto]

You're assuming that in 10 years they'll have made no optimizations to python?

Is that such an unreasonable assumption to have? There are plenty of ways to make python dramatically faster. None of them has happened so far, with excuses ranging from backwards compatibility, to user-friendliness, cleanness of code, or too much work. If you want a faster python, don't hold your breath, instead either switch language, or just be happy that even if python isn't particularly fast, at least it is particularly nice.

but even the currently less popular ruby is getting more than twice as fast at basically everything in the next version.

But Ruby is also slower than python, so all the obvious optimizations aren't all done yet. Besides, it's always dangerous to talk about the "next version". Where's parrot (for perl)?

No, Python is still going to be slow. And so is Ruby. A 50% or 100% increase in speed doesn't really matter much, when it's already orders of magnitude slower than C. That isn't necessarily bad, or even something that needs excuses. Throughout the history of computing (well, since the invention of Fortran and LISP), there has always been a market for nice slow languages, as well as ugly fast languages. This isn't unique to 2007, or 2017, and we don't need to invent reasons for slow languages to exist, such as concurrency (although there are slow languages invented with concurrency in mind too). Slow languages exist because they are useful. People will always keep inventing useful languages that are slow, and people will always continue to use something if it's useful, even if it's slow!

Re:Python

[the_lesser_gatsby]

Python should reinvent itself to be the multi-processing language

Why bother when Java has done that successfully for years? (And with performance on a par with your unoptimized C).

Re:Python

[vtcodger]

***But Ruby is also slower than python, so all the obvious optimizations aren't all done yet. Besides, it's always dangerous to talk about the "next version". Where's parrot (for perl)? ***

The next version of Python is 3.0/3000 and the first Alpha was released three weeks ago. Scheduled final release is August 2008. I don't believe that it is dramatically faster or has any breakthrough in multicore support, but I'm no expert and I'm often a bit baffled about whether specific Python improvements are a big deal or not. FWIW, here's the wikipedia link http://en.wikipedia.org/wiki/Python_3

Re:Python

[philipgar]

IN ten years, according to moore's law python will be 32 times faster than it is now.

Moore's law says nothing at all about the speed of a processor, or of a program. It only says the number of transistors will double every 2 years. The fact that performance benefits have traditionally been had by adding transistors does not mean this will hold. In fact today the performance of most applications is no faster on current computers than top of the line computers from 2 years ago (it's definitely not twice).

Phil

Re:Python

[try_anything]

Java did not go far from the plain old mutex, lock, and condition variable paradigm. Java's version is almost as flexible and a lot easier for beginners to understand, and Java steers beginners away from incorrectness (towards slowness, but that's the right direction for beginners.) It's still not so different from "raw" threading using POSIX, just more elegantly expressed. Outside Java, everyone is still doing "raw" threading in all of the popular languages, while various less-popular languages such as Erlang and Scala(*) experiment with radically different approaches. When the hammer starts to come down and more and more J. Random Codemonkeys have to take advantage of multicore processors or lose their jobs, there will be a big migration to one or more languages that make threading as easy as possible. I think that's what he's talking about when he says Python should be "the multi-processing language."

(*) People are stubborn in their prejudices and have already categorized Erlang and Scala as being "too hard" or "weird" or "languages for smart people." Even if Erlang, Scala, or another language has already nailed the next generation of threading concepts, people will wait for them to appear in a language with comfortable, low-brow marketing that assures them they won't fail to make the leap. The only thing Python has going against it is that many very smart people like Python -- HPC scientists use it to wrangle Fortran libraries, ESR famously praised it, it's popular in Google for scripting, and IIRC even Peter Norvig has said some kind words about it. Hopefully that doesn't doom it, because most of those people praise it for being less intellectually taxing than its similarly-powerful competitors.

OT: Thin processors

[Anonymous Coward]

My favorite MacGyver episode is where he visited a college. One student shouts, "I got my processor design down to 4 atoms thick!"

What made it so funny is that they made it clear that these were undergraduates. Don't forget that this was in the eighties, too.

Simple Solution

[Nom du Keyboard]

the number of transistors on integrated circuits would double every two years.

The solution is simple. Just make integrated circuit dies twice as big every two years.

Avast!

[ackthpt]

Exactly. Whenever one process technology reaches its physical limits, we get a new one, because the new process makes money. X-ray lithography, chip stacking, 3D circuits, and eventually nanotech will all keep us on the Moore's law path probably for the rest of my life, at least.

Ye be forgettin' one thing, matey, they be makin' multiple cores now. Eventually we be lookin at distributed computing on an individual platform. Ye may be layin' claim to Moore's law applyin', but it be tenuous a claim at best. The paradigm be shiftin' away from the domain of Moore.

Re:Avast!

[MightyYar]

Why do you think that multiple cores and Moore's Law mutually exclusive? All Moore's law does is predict that the number of transistors on a given chunk of silicon will increase exponentially. Put 1 or 80 cores on the same chunk of silicon... it makes no difference as far as Moore's law is concerned.

Oh yeah, yaaaaarrrr.

Re:Avast!

[ackthpt]

Why do you think that multiple cores and Moore's Law mutually exclusive? All Moore's law does is predict that the number of transistors on a given chunk of silicon will increase exponentially. Put 1 or 80 cores on the same chunk of silicon... it makes no difference as far as Moore's law is concerned.

Oh yeah, yaaaaarrrr.

Arr! Intel be using separate dies to accomplish multiple cores. At some point even AMD be havin' to, matey. Emphasis be movin' to lower cost production, just be splicin' the mainbrace, er, splicin' multiple inexpensive dies to be accomodatin' more cores they be!

Re:Avast!

[MightyYar]

Wow, they are really using separate dies now? I'm surprised - they got themselves into some trouble back in the day with the Pentium Pro when they used separate chips for the processor and cache. It's hard to test both chips, and so you end up assembling everything with the potential of losing the whole package if final test fails. It also would make inter-core communication slow because of the much larger wire distances.

Re:Avast!

[ackthpt]

Wow, they are really using separate dies now? I'm surprised - they got themselves into some trouble back in the day with the Pentium Pro when they used separate chips for the processor and cache. It's hard to test both chips, and so you end up assembling everything with the potential of losing the whole package if final test fails. It also would make inter-core communication slow because of the much larger wire distances.

Aye, they be taking risks and writin' off some speed, but they must have some secret Intel way of testin' things. AMD be staying with one die, for now, but processors be commodities now, ye be throwing many at the problem and less emphasis on how potent each one be. Aye be thinkin' that be how Google got started, just on different a scale.

Re:Avast!

[MightyYar]

Yikes, I just found some pictures o' it - guess I haven't been keepin' up since I started workin' from home. That thar be an ugly hack - but probably a godsend for anyone doin' video or renderin' work. By definition, that package be goin' t' be slightly more than 2x t' cost o' a sin'le-die solution (assumin' t' die be t' same size). Still has t' be cheaper than puttin' all four on one big die - plus you can reuse t' dual-core fab line.

Mod up

[Anonymous Coward]

Post is not off topic.
Gains in CPU speed affect the utility of scripting languages more than compiled languages. Parent makes this connection is a humorous way.

End of Moore's law!

[Deadplant]

The end of Moore's law! New solar panels with double efficiency! Flying cars now only 5 years away!!!!

Are these articles being generated by a script or what?

10 years ...

[jbeaupre]

Pretty much what he is saying now. So a corollary might be that when Moore stops predicting, Moore's law only has 10 years to run. Which means we all better hope he doesn't die any time soon.

nonsense

[Quadraginta]

That's nonsense. The industry grew around the physics, not vice versa. The fact that the industry is predicated on a constant improvement of speed and complexity is because such a thing is achievable in microelectronics, certainly not because microelectronics is the only industry where such a thing is desirable.

I mean, who wouldn't want cars to become twice as gas efficient (without losing power) every 18 months, ad infinitum? If such a thing were technically possible, it would happen, because all the car makers would jump on the gas-mileage bandwagon to get ahead of their competitors.

Who wouldn't want the amount of food that can be grown per man-hour to double every 18 months, so the price per pound of beans and broccoli fell as fast as the price per CPU cycle of computers? If such a thing were possible, it would happen, as every farmer raced to lower his costs of production and undersell his neighbors like crazy, earning millions.

In very few industries other than microelectronics has anything like Moore's Law applied, and that's not from a lack of economic incentive, but from the plain uncooperativity of Mother Nature. You're arguing backwards, from effect (the economic structure of the industry) to cause (the physical nature of microelectronics).

Re:nonsense

[avirrey]

I mean, who wouldn't want cars to become twice as gas efficient (without losing power) every 18 months, ad infinitum? If such a thing were technically possible, it would happen, because all the car makers would jump on the gas-mileage bandwagon to get ahead of their competitors.

Hah! Exxon wouldn't, and they don't, so we have no say.
--
X's and O's for all my foes.

Re:nonsense

[klenwell]

I mean, who wouldn't want cars to become twice as gas efficient (without losing power) every 18 months, ad infinitum? If such a thing were technically possible, it would happen, because all the car makers would jump on the gas-mileage bandwagon to get ahead of their competitors.

An international cartel of oil/gas producers perhaps? An auto industry focused on short-term profits at the expense of long-term economic and environmental (which is effectively the same thing) viability? A corrupt state power beholden to lobbyists of the previous two groups.

Who wouldn't want the amount of food that can be grown per man-hour to double every 18 months, so the price per pound of beans and broccoli fell as fast as the price per CPU cycle of computers? If such a thing were possible, it would happen, as every farmer raced to lower his costs of production and undersell his neighbors like crazy, earning millions.

An established cartel of successful food producers? An agro-industrial complex focused on, &c.

To bring a popular example around here, who wouldn't want a better, more secure operating system at a cheaper price?

Why is the microelectronics industry not subject to the same market manipulation by established powers as other industries? The answer is obviously complex, but an important factor seems to be the fact that it has not yet descended into the sort of cartel-ishness that seems to be the natural tendency in completely unregulated markets.

I guess I agree with your answer -- the physics is a big part. But your analogies don't really make sense to me. And I don't think any new industry matures into a stable business sector around the physics. If does so around the market and the type of market the state shapes.

If you eliminated the economic incentive -- some minimal level of competition -- I wouldn't expect Moore's Law to magically continue because Mother Nature makes it possible.

Re:nonsense

[Anonymous Coward]

Where do all you fucking conspiracy theory locos come from, and why are you all on slashdot?

Re:nonsense

[Kirijini]

Others have jumped on your examples, and I will too. Fact of the matter is that there are political reasons why cars can't become twice as efficient, and political reasons why farming can't double in efficiency as well - in addition to whatever limits laws of nature, natural resources, etc put on it.

If food efficiency doubled every two years, there would be a glut of food, and prices would drop drastically. This actually happened in the early 20th century, and its why we have agricultural subsidies, and price floors.

What it comes down to, I think, is demand. There is a constant demand for more computational power, but not so for food (people don't starve for lack of food, but lack of money). While gas prices are low (admit it, even 5 dollars a gallon is pretty cheap for the amount of power you get out of it), there's little incentive to invest in gas efficiency. But, you may have noticed the trend of increasing gas efficiency (but not doubling) with increasing gas costs. Just goes to show that demand seems to be the driving force.

Re:nonsense

[hawk]

>I mean, who wouldn't want cars to become twice as gas efficient (without losing power) every 18 months, ad infinitum?

Once it was efficient enough that fueling was only as frequent as other changes (e.g., oil change), further improvements would be insignificant.

hawk

Re:nonsense

[servognome]

An international cartel of oil/gas producers perhaps? An auto industry focused on short-term profits at the expense of long-term economic and environmental (which is effectively the same thing) viability? A corrupt state power beholden to lobbyists of the previous two groups.

Or more likely consumers who were more interested in horsepower, 0-60, and low cost rather than long-term environmental impact.

An established cartel of successful food producers? An agro-industrial complex focused on, &c.

You mean the same industry that is investing in geneticly modifying food to make it cheaper, faster growing, etc.

To bring a popular example around here, who wouldn't want a better, more secure operating system at a cheaper price?

People who are more interested in an operating system that runs the applications they need.

Re:nonsense

[suv4x4]

In very few industries other than microelectronics has anything like Moore's Law applied, and that's not from a lack of economic incentive, but from the plain uncooperativity of Mother Nature.

Well, rather microelectronics is the youngest technology of everything you mentioned, still in development. There was a time when speed and fuel efficiency doubled each 18 months for cars too.

Moore's Law isn't eternal though. Even worse, it's not eternally relevant. Remember P4-s? They had more transistors than Pentium M but ran slower and ate more electricity.

Every year it's Moore's law last year

[Wiseman1024]

Moore's law will end the year of the Linux desktop; the same year Duke Nukem Forever, Parrot, Perl 6 and bytecode compiled Ruby will be released.

sure they do

[Quadraginta]

They only change in ways that are generally not possible to anticipate, hence which haven't been predicted.

And of course they would. Technology, like the stock market or the weather, is inherently a chaotic system over a certain characteristic timespan (1-2 weeks for the stock market and the weather, 25-50 years for technology). That is, over the characteristic timespan very small causes can produce enormous, system-wide effects, what you might call the butterfly wing flapping causing the hurricane phenomenon.

For example, a couple of guys (Jobs and Wozniak) screw around in the garage in the early 80s, trying to put together a really cheap personal computer. That's a very small cause. And twenty-five years later, it has had a giant effect: iMacs and iPods and iTunes oh my. Problem is, there was no practical way in the 1980s to distinguish the small cause that mattered (Jobs and Wozniak) from the other 50 zillion small causes that didn't matter (the other 50 zillion pairs of scruffy entrepreneurs in garages whose brilliant idea went nowhere).

This is why predictions of the future out more than 50 years usually end up looking hilarious in hindsight. When sf writers of the 50s and 60s predicted the present, they projected the dominant themes of their time (spaceflight, atomic physics, the struggle with Soviet Communism). They did not -- and could not -- realize that all three themes would pretty much abruptly and surprisingly come to an end in the 90s. When present writers predict the future, they project the dominant themes of our times (e.g. networked computing). It's very likely these projections, too, will end up wildly wrong. Networked computing is likely to become as humdrum and static as telephony within the next half-century or so.

Didn't I hear about this...

[TheDarkener]

about 10-15 years ago?

Do you ever notice?

[Seismologist]

Do you ever notice that articles or post beginning with a question in the leader are crap?

Canned food and bottled water.

[mcmonkey]

While we all tatter on about whether Moore's Law concerns economical or physical constraints and come up with cure meta-laws concerning predictions to the end of Moore's Law, have any of you really considered what the world will be like after the end of Moore's Law?

Are any of us prepared for Moore's Anarchy which will surely follow?

the terrorism thread advisory level

[game kid]

Yesterday, Bush told Americans to continue with their login sessions, but to look out for suspicious threads. Hackers worldwide, meanwhile, called Bush's anti-terrorist-thread practices--both on the desktop and over the network--"thinly-veiled discrimination" against O(c^n) processes, and plan to stage a general protection fault on September 11.

When asked for comment, Intel simply sent a letter describing how their latest Core 2 Duo could do an infinite loop in just 3 seconds, instead of 5 like "that free operating system thingy".

Moore's Law has nothing to do with CPU clock freq

[cciRRus]

From Wikipedia:

Moore's Law describes an important trend in the history of computer hardware: that the number of transistors that can be inexpensively placed on an integrated circuit is increasing exponentially, doubling approximately every two years.

It has to do with the number of transistors and not the operating frequency.

Yeah right

[Quadraginta]

so we have no say.

Who's this "we" of which you speak? Do you mean you personally and the dozen people you know personally? In that case, you're right, you have no perceptible influence on a corporation employing tens of thousands and providing a service for millions. Nor should you. I'd object very strongly if you did, since I have no use for an all-powerful aristocracy of any kind.

Or by "we" do you mean "all 50 million of Exxon's customers?" In which case, it's obvious to anyone using a human brain instead of, say, a 50-line Perl script loaded up with 50 megabytes of mindless slogans from the past that that "we" has enormous influence. All that "we" needs to do is switch from Exxon gas to BP or Mobil gas for, say, six months or so, and Exxon would be totally ruined, driven completely out of business. Indeed, a corporation that large is so dependent on a continuous stream of income that I expect a single day during which no one bought their gas would be so horrifying to upper management that they'd do nearly anything to avoid it.

I suspect what you mean by "we" is something between these two extremes. A "we" that obviously does not include all customers of Exxon, but is somewhat larger than you and your personal friends. Say, you, your personal friends, and all the other clever in-the-know people. It's still pretty OK with me that this particular "we" has just about zero influence on Exxon, since it won't include me and my personal friends.

Re:Yeah right

[avirrey]

Any perception of symbolism?

"Exxon" as used is symbolic.
"we" as used is symbolic.
Think beyond your literal shell if you could.
And I'm curious on how "you" and "your friends" who are excluded from the inability to "influence" Exxon is. Do you walk to work? Good for you.

I'm surprised people get modded-up now a days for exessive use of the word "we".
--
X's and O's for all my foes.

Re:Yeah right

[Quadraginta]

Sure. But I read "symbolism" in this context as "license to bullshit vaguely without defining your terms in a way that might leave you open to counter-argument." Useful stuff in a class on literary criticism for English majors, so everyone can be right in their own particular way. Totally worthless to someone with a more scientific or engineering frame of mind, who wants to pin stuff down so that it's measurably true or false. Since I'm not a lawyer or English major, I consider this kind of symbolism to be worthless.

Think beyond your literal shell if you could.

Why? So I can appreciate the Cosmic Oneness of it All? Be more open to Ultimate Truths that are too truthy for mere words and grubby logic? No thanks. Tried smoking that stuff once in college, found out it interferes with being able to solve differential equations and earn a good salary.

And I'm curious on how "you" and "your friends" who are excluded from the inability to "influence" Exxon is. Do you walk to work? Good for you.

You've misunderstood me. In this world, the way it actually is, I have influence over Exxon, the same as any other customer. If I don't like their product, or their behavior, I can not buy their gas. They can't force me to buy it. However, in the fantasy world that the OP would like to live in, where his "we" has great influence over Exxon, I would lose the small influence I presently have. Because his "we" wouldn't include me. So basically he and his friends get to determine what Exxon does, and, therefore, what kind of gas or other energy supplies I get to buy. I'm not in favor of that. I like my freedom.

Moore's 2nd Law

[Trailer Trash]

Moore's 2nd law is that Moore's 1st law is going to come to an end in about 10 years. Always.

uh yeah

[Quadraginta]

So you're saying all human economic endeavors grow exponentially until they can't?

Amazing! Who'd have thought?

Roman's Law

[roman_mir]

Roman's Law states that every year someone will predict the end of Moore's Law and they will be wrong.

Wow..

[way2trivial]

I never knew EXXON manufactured CARS

Re:Wow..

[avirrey]

Totally. I never new either until I started working in the automotive industry. =)

My paycheck says "Exxon" on it. LOL.

--
X's and O's for all my foes.

Like Anybody Knows What Will Be in 10-15 years

[littlewink]

We can't get a good economic forecast six months out!

10 years?? Fugetaboutit!

Actually Moore's Law is common

[Anonymous Coward]

As The Innovator's Dilemma documents, in many industries where there is a clear criteria of what is "better", there tends to be exponential improvement in that metric over periods of decades What is unusual about Moore's law is the speed of the exponential improvement, not the fact of it.

And this has been true whether that metric was the distance a steam ship could travel, the volume of dirt an excavator can pick up, or the quality of steel a mini-mill can produce.

Moore's Law as Energizer Bunny: not about silicon.

[mschuyler]

Moore is being short-sighted about his own law. It's not about silicon. If you extraploate backwards from the first integrated chip you see that "Moore's Law" has been in effect for over 100 years. It started with manual switches, then moved to electric motor switching, then to vacuum tubes, then to transistors, then to integrated circuits. Every one of those mediums has been subject to and demonstrates Moore's Law. Graph it and you'll see. It's a perfect logarithmic line. Every time the method itself peaks of its own accord a new medium is found which can continue the progress. (Any familiar with the growth of telco equipment can see this in the switching systems: Electric switches to step systems to crossbar to ESS.) If IC does run out, there is a future of possibilities: holographic, quantum, bio, etc. Moore's Law is like the Energizer Bunny. It just keeps going.

Need another law

[slapout]

Someone should make a law about how often people speculate how long Moore's law will last.

Moore's "End of Moore's Law" Law

[WED Fan]

As we approach the stated end date, a new end date will be assumed to be stated as another 10 to 15 years in the future. (This is also known as the "Televangelist End of the World the Rapture is Around the Corner Law" (TEWRAC).)

Less's Moore Law

[CaptDeuce]

Moore has predicted the end of his own law in 10 to 15 years

Less's Moore Law predicts that the predicted rate of prediction production of the predicted end of Moore's Law increases by 1.15 times per month as the predicted end of Moore's Law approaches.

Note: if my name was actually Les, then the name of Less's Moore Law would be different -- and totally awsome.

Moore's Law is a DEAD parrot

[RandCraw]

Moore's Law describes a CPU speedup that died at least 3 years ago (all other legalisms aside).

To wit: I bought a laptop in 2003 with a 2.2 GHz 32 bit P4. According to The Law, by 2005 CPUs on comparable laptops should have run at 4.4 GHz, and by today they should zip along at 8.8 GHz. But in fact, no commodity CPU runs at that speed nor even *half* that speed.

And don't you believe the claim that multicore or power throttling compensates for or explains The Law's "failure to thrive". The fact is, the industry is no longer delivering CPUs whose SPECMarks/FLOPs/etc (AKA performance) is rising at the rate that they have for the previous 20 years. I tell you, "Moore's Law is pushin' up daisies. It's a DEAD parrot."

What's puzzling to me is that while this emperor is clearly naked, for some reason, nobody wants to admit it. Why not? Are we afraid that sexy soothsayers like Ray Kurzweil or Rod Brooks will be regarded laughably when they forsee cool stuff like The Singularity or robots possessing human-level cognition, brought to us by the inexorable exponential march of Moore's Law? Or do we simply dread the day when we have to depend entirely on advances in *software* to deliver our next high-tech fix? Perish forbid *that* thought.

Well, we'd better get used to it, the emperor is naked *and* dead. There's a new emperor in town, and Moore's Law 2.0 depicts a future that looks a hell of a lot like the past.

        Randy

Re:Moore's Law is a DEAD parrot

[jgowen]

I agree; CPUs have *not* increased in speed as per Moore's Observation for some years. I would suggest that the constant puffery for multi-cores and parallel processing is *evidence* of that. And I think Microsoft at least knows and fears it; see the upcoming PLINQ feature for C# etc. (http://home.att.net/~owen_labs/rant14.htm#revealed).

Re:Moore's Law is a DEAD parrot

[DragonWriter]

CPUs have *not* increased in speed as per Moore's Observation for some years.

Moore's Law isn't CPU speeds, its about transistor densities. Its been informally extended by lots of other people to "computing power" more generally and all kinds of other particular measures of computing power besides transistor densities, but those aren't "Moore's Law" but loose parallels of Moore's Law.

Anyhow, "clock speed" and "CPU speed" aren't really the same thing; more advanced CPUs with the same clock speed can do more operations, on average, per clock tick, and thus be faster than less advanced CPUs of the same speed. The death of the GHz race hasn't been the death of performance improvements, its just that ratcheting up clock speeds has turned out to no longer be the easiest way to squeeze more performance out of a chip, so that's not the main focus currently.

Moore's Law end will be lack of interest

[chipace]

I don't think that technological barriers will bring us to the end of Moore's Law (predicted in 10-15 years)... but rather consumer interest. I have always spent more money on my display, than on the box that drives it. I now spend more on storage (hard disks and blank dvd media) than on silicon (be that cpus, motherboards, dram, flash, etc...).

I believe that I'm at the point where replacement is the only way I will spend more money on silicon (this is also true about my displays). Hard disk storage still has good value (every two years they double the amount at a given price point)... and I am always accumulating personal/entertainment pictures & video.

Am I the only one disappointed with the functionality of new silicon?

Says who?

[Estanislao Mart�nez]

You're talking as if philosophy of science were not an insanely controversial field that fails to agree on stuff seemingly as basic as that.

your examples don't hold up

[goombah99]

Youre counter examples are actually arguments in favor of what I said.

The price of food production DID fall exponentially over many years. Then as I predicted when margins approach development costs, it became a commodity without any quick doubling time. The cost of oil production did fall exponentially over many years then it became a commodity.

Moreover you are comparing resource extraction with technology development. No matter how hard you try there simply is only so much energy in a barrel of oil. once you get within a factor of 4 or less there's not much doubling left to get. Likewise once transportation costs exceed the production costs of brocolli, the problem is not in the same realm.

If a way is found around Einsteinian physics....

[Anonymous Coward]

restrictions concerning the speed of light it won't be used to explore
Alpha Centauri, it will be to keep pushing Moore's law and build
faster processors, and the first of those will create the singularity
because "this new CPU really sucksssssssss"

Recursive Moore's law

[chudnall]

"The number of people predicting the end of Moore's law will double every year"

And lg 10 to lg 15 is ...

[crmartin]

3.32 to 3.91 ... so according to Moore we can only expect another 3 or 4 doublings.

Oh, no! What will we ever do!

Moore's Law: addendum

[Kingrames]

Moore's Law will end, when the following occurs:

When 2^n+1 is no longer greater than 2^n.

That is to say, when you no longer gain something from doubling a number.

Past performance is no guarantee of future results

[turing_m]

It may do that, or it may not, much like speed for commercial travel has hit the point of diminishing returns. Processor speed has been remarkably similar - you get the overstep that results in products like the Concorde and Williamette P4 - moderately faster, much more expensive. Then you get a move to more efficiency while maintaining similar performance. Of course, that depends on what the physics is. If increasing flight time by 25% shaved 90% of the fuel cost, jets may have gone down that path.

Sounds About Right

[YetAnotherBob]

IC's today are made photographically, on a flat surface. Manufacturers keep working to reduce the area needed for a component, be it transistor, resistor, capacitor or trace wire. We already know from lab work what the minimum possible sizes are for each basic component. We've come up on the minimum possible size several times in the past. Each time, it was related to the possibilities of the light source we were using. Now, we are up there in the extreme UV range, and have minimum feature sizes that are actually smaller than the wavelength used. The best commercial plants use a 45 nM wavelength. At about 30 nM, the traces (on chip wires) become unstable, and may no longer be conductors. That is a fundamental limit that clever plant engineering will not be able to surmount. Current commercial plants are using a 60 to 90 nM min. feature size, if memory serves. That means we have about 6 or 7 doublings (each doubling is about a 70% reduction in feature size and takes 2 to 3 years t realize.) That gives us 12 to 20 years.

Going to still smaller wavelengths means that the photons pack more punch. It's like trying to play billiards by shooting the cue ball with a high powered rifle. You get pieces of cue ball everywhere. When random photon collisions are pushing random atoms by several dozen radii, your nice ordered atomic lattice becomes a horrid mess. we are nearing the limits of what nature allows for photo lithography now.

Increasing chip size is not a viable solution, as the full wafer is used now. Increase chip size, and yield drops quickly. Yes, they could double the size of the chip to increase transistor count, but that would mean increasing the cost of the chip by 4X. That's not he direction we want chip cost to go.

Off in the distance, there are more real hard boundaries, beyond which no amount of effort will yield additional benefits. One of those is component size. Minimum transistor size is 7 atoms (it's been done). Minimum diode size is about 5. Minimum trace size varies with material. The best I've seen is benzene, at about 6 atoms width. Keep in mind that at room temperature, benzene is a gas. It's going to be very hard to make wires of the stuff. We really need a solid. Aluminum, silver, gold, all have been used, and all need to be 30 to 60 atoms wide or more, and several thick to be even a poor conductor. Some creative metallo-insulator engineered materials might allow for smaller trace sizes, but probably not. Please note that this is still smaller than buckytubes, which are also as tall as they are wide, creating other connection problems, so don't peddle that as a panacea. That means that the trace sizes required will probably be the final limit. Real capacitors are larger than the traces, but their size is really controlled by the number of electrons needed to operate the transistor/switch. I'm still betting on the traces as establishing the limit.

Heat dissipation is also a problem. It gets to be more of a problem as densities go up. Current best designs are operating half way to melt now. switching to silicon carbide would let us go hotter, say 400 to 800 C. Diamond/graphite bases would let it get higher still, though diamond heated to 1,200 in an oxygen atmosphere isn't going to last very long. Need some creative packaging there. Heat dissipation is the real reason we can't go 3D. The systems that tried to be true 3D, or near to it, all relied on the chips being immersed in some coolant and having channels for the coolant through the chip. Liquid nitrogen cooled some that IBM did a few years ago. bubbles were a problem. move the coolant fast enough to transport the heat before bubbling and erosion is a problem.

Some of these issues can be fixed, some can never be fixed. So, when we are fully 30 nM size with our components, it all stops. It's a problem with the wiring. Solve that, and we would be close to being able to compute with atoms. But, with what we think we can do now, the shrinkage stops in about 20 years.

Enjoy it while you can.

Looks like you

Moore's Law? Pfft.

[boarder8925]

I prefer Godwin's Law myself.

The only law that is timeless is....

[Aussie Osbourne]

Cole's law: thinly sliced cabbage.

/ducks

Better software may surpass hardware

[OricAtmos48K]

A minimum of 15 years of growth may mean 2^10 more raw power. But I guess using the raw power with better designed software would make a lot of difference 30 years later. Assuming that we will reach a hardware limit

Re:Moore's Law has nothing to do with CPU clock fr

[dylan_-]

I know. I have heard of it before ;-)

  My point was more a general one on predicting limitations on technology.

Did you read the first five words?

[benhocking]

It was an explicit acknowledgment that I was simplifying the matter. Nevertheless, it's a reasonable simplification of the difference between a law and a theory.

Re:Past performance is no guarantee of future resu

[mschuyler]

I agree. The real point is that Moore's Law is not dependent on Moore, nor on silicon. If in the past researchers had fixated on the vacuum tube, they never would have reached beyond the vacuum tube paradigm to make the advances that happened. I am encouraged by the results other research labs have already achieved with these new mediums. It's not so much that they still need to be invented as much as it is that their discoveries need to be developed. I think it was William Gibson who said, "The future is already here. It's just unevenly distributed."