Slashdot Mirror
The Mile Markers of Moore's Law Are Meaningless
szotz writes "Keeping up the pace of Moore's Law is hard, but you wouldn't know it from the way chipmakers name their technology. The semiconductor industry's names for chip generations (Intel's 22nm, TSMC's 28nm, etc) have very little to do with actual physical sizes, says IEEE Spectrum. And the disconnect is only getting bigger. For the first time, the "pay us to make your chip" foundries are offering a new process (with a smaller-sounding name) that will produce chips that are no denser than their forbears. The move is not a popular one."
Is that some archaic form of measurement used by a backwards nation somewhere?
-- Tigger warning: This post may contain tiggers! --
And thirdly, More's law is more what you'd call "guidelines" than actual rules.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
Moore's Law isn't even a law... it's a prediction. Didn't we already agree that predictions are only useful to talking heads, pundits and hucksters?
Anyone who actually works in the semiconductor industry could've told you this. (Ever notice how the GHz stopped growing a while ago? The move to multi-core happened around the same time and even that's stopped growing.) Yes, it's still possible to shrink transistors further but the speed and power reduction gains are diminishing and the costs of further shrinking are moving from merely eye-popping to astronomical.
Intel can afford to stay ahead of everyone else a bit (this is one of the primary reasons AMD is having difficulty staying alive) because of the huge volume that they have but even they're having problems.
all are antiquated.
Surely you jest!
That way, you'd have the option of scrolling back to less dense chip designs.
Changing the names to make something sound better has been a strat for decades, if not longer.
So why is this a surprise that the semiconductors are using it now to sell stuff.
Be seeing you...
If TSMC isn't keeping up with Moore's Law, that's not a problem with Moore's Law. It's a problem with TSMC.
>
Waaaay towards the end of TFA, it mentions that it's GlobalFoundries who inserted finFETs into the same BEOL (wiring) as their 22nm node and called 22nm+finFET "14nm." It's buried at the end, but it's what supports the whole argument that nodes are "just marketing."
To my knowledge, the node's name was based on the DRAM half pitch. But yeah, it's not that any longer. And in defense of GlobalFoundries, finFET does literally add an extra dimension to the calculation of FET geometries.
Who do you get to be an expert to tell you something's not obvious? The least insightful person you can find? -J Roberts
You've got it switched...
see, when the data does not support the hypothesis, you **change the hypothesis** not how you interpret the data
Moore's Law has never been a 'law'...it was a cool statistical novelty that seemed to predict processor advancements...it is NOT and HAS NEVER BEEN fit to predict anything invovling money or resources...it's 'for fun'
I've seen Singularity/Kurzweil types in TED talks show some dumb graph of 'Moore's Law' and show how, according to the law, humans will have the processor speed to do XYZ by 2050....it's all bunk...
Using Moore's Law to make important decisions is about like using a Slashdot Poll to do the same...I don't trust people professionally who take a concept like Moore's Law and build their understanding of an industry around it. It's a common mistake of perception.
Maybe there is some sort of pattern to processor speed, but it's not helping us understand anything to be so reductive and irresponsible with how we use scientific concepts.
Thank you Dave Raggett
The problem with the transitions to finFETs is now we have an apples-to-oranges comparison between finFET (or 3D gate or whatever you want to call it) processes and surface FET processes. GlobalFoundries feels they need to stretch the truth to get the point across that the process really is better objectively, even if the minimum feature size hasn't shrunk.
It reminds me of 10 years ago when the microprocessor companies finally stopped the GHz war. For several years, clock speed was a poor proxy for microprocessor performance, and Mac fans used to scream loudly (and rightly) how the IBM chips beat Intel on real-world benchmarks while Intel touted their higher speed.
Hopefully, this "node as minimum gate width" will go away and we'll move to more meaningful process figures-of-merit such as power density, power-delay product, gm/I, transit frequency, Ioff and the like.
In the middle of the article it is pointing out that Intels 22nm chips use gates that are 35nm long with channel lengths that are 30nm long, so it seems odd that people are worried about what GlobalFoundaries will be misnaming their not-yet-in-production 3D chips when Intel is already misnaming their already-in-production 3D chips.
"His name was James Damore."
"...and Mac fans used to scream loudly (and rightly) how the IBM chips beat Intel on real-world benchmarks while Intel touted their higher speed."
Mac fans used to scream loudly about anything that made Macs look good...and still do. It's called tribalism and it isn't about being "right", it's about being on the winning team.
Apple only used an IBM "chip" once. It's clear you don't know that so it's no surprise you don't know how "rightly" Mac fans were about their screaming either. G5's were, in the balance, not faster than their Intel contemporaries. Better at some things and worse at others. One thing was clear with the G5 and it was that Apple was switching to Intel afterward.
If you asked any "Mac fan" back in the day you'd get explained to you just how superior every generation of PowerPC Mac was to any PC ever. It's surprising then, just how much better Macs got once they switched to a real processor. Macs today ARE PCs in every way yet those Mac fans still have that feeling of smug superiority. They are inherently right always, Steve told them so, they just aren't well informed.
Mac fans used to scream loudly (and rightly) how the IBM chips beat Intel...
Umm, don't you mean Motorola chips? I mean, I'm no Appel expert but I could have sworn appel was big on Motorola procs for the longest time. I always liked the flat model addressing in those...
Python: 'And then suddenly you have a language which says "we're all stuck with whatever the whiniest coder wants".'
What a Hitlerian thing to say.
Python: 'And then suddenly you have a language which says "we're all stuck with whatever the whiniest coder wants".'
hey thanks for the response
Right, so did you just use Moore's Law or did you look at other factors as well?
What I mean by other factors:
> Trends of the capacity of other recent products? Did you look at teh speeds of CMOS processes from that company over the last 10 years and extrapolate?
> Did you talk to a sales rep or engineer or product development manager at the CMOS process company and **ASK THEM** how fast their upcomming models would be (approximately)
> Do literature review of what academic research groups and possible FOSS (idk if it applies for you) were doing in that CMOS wireless type transciever tech? My former university, Ball State University did research for WiMax coverage and speed for Cisco (before WiMax was ditched)...did you look at any of that to predict the CMOS process capability you needed?
I'm trying to be polite, but I call BS.
If you claim your company made that decision based **soley** on math from Moore's Law....well I have a hard time believe that claim's veracity. You are either fabricating or that company is not very wise. And if you company **did** use other factors, then that kind of invalidates your point and parenthetically supportsy my point...I won't deny that using it **might** have added value, but only IF you also did common practices like I mentioned above...
Seriously...did you use other factors besides Moore's Law?
Like asking the vendor? (or any of the others mentioned above)
Thank you Dave Raggett
G3 and G4 were Macs using IBM designed (largely) processors. Motorola and IBM jointly produced Power PC chips that Apple used in the mid/late 90s (G3 and G4) but Motorola eventually dropped out and IBM wasn't interested in keeping up with Intel. For a few years, the Apple chips were better than the IBM chips (I didn't own an Apple computer at the time, so I was evaluating this as an engineer). By the time Intel had closed the gap Apple wisely went over to the Intel architecture.
Even if we did, there are not enough electrons in these lines to make the "law of large numbers" work. So this time we are bumping against a real barrier.
Anyway, there are not any mass market killer apps based on computation anymore. All the action is in connectivity and bandwidth enhancement. Given the computer market has been split into makers vs takers (or content produces vs content consumers) this is changing the funding models. Earlier the large number of passive consumers buying computers way more powerful computationally than what the typical consumer needs, was subsidizing the cost of computers for the few who actually need that much of computational power. Now the passive consumers are buying simpler devices needing less computation and more connectivity. We can expect coders like us can expect our hardware to get more expensive, like the old line of unix workstations like micro-vaxes or sun-solaris or hp-ux or SG-Iris.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
And yet the channel width is ~8 nm, which is ~64 atomic layers. How many times do you think they can cut that in half? And does it really matter when the source and drain contacts are 10x the size of the channel itself?
All sub-65nm and most 65nm processes are lithographically exposed in water current for the reason you stated. The next step is extreme UV or even e-beam lithography but it's expensive and very, very difficult.
You're quite right that this is an economic/mass-market issue more than a pure technical issue.
For the case of comparing GPU and CPU, that's a pretty good assumption, because they are both digital logic dominated. Once you start mixing in other functions, things can change dramatically.
I'm not sure where you heard that "20nm is the first process node which will be more expensive than the preceding one". I'm curious how that could be argued. Nearly every tool has grown near exponentially in price. Perhaps by scaling up production they've been able to keep reducing cost, though.
thnx again...
so, go up a few branches and you'll find your original comment...it **didn't mention other factors** and, most importantly, you said that the success was **due to using Moore's Law**
indep. of each other, fine, but you used pretty flowery language to describe the pressures of your decision and cited **only** Moore's Law for your making the right choice...here's one example:
but throughout you only attributed your success to Moores Law...
Other commenter here on this branch is right...we don't need to dance around the issue...Moore's Law is an interesting novelty and that's all...nothing wrong with running the numbers on it for comparison sake (b/c others in the industry use it if nothing else!)
Thank you Dave Raggett
5 years, eh? You just missed out on the previous generation workstation that experienced the coolant leak debacle, where your Powermac G5 would suddenly leak the coolant they were using down over the motherboard and power supply and then out the bottom of the chassis.
Heat is a big cpu problem. Have they tried rotating the active core on the fly? When core 0 gets too hot, they switch it off (under single core workload of course) and move activity to core 1. When that gets too hot, I type a bunch more, or you get the picture.
I come here for the love
To my knowledge, the node's name was based on the DRAM half pitch. But yeah, it's not that any longer. And in defense of GlobalFoundries, finFET does literally add an extra dimension to the calculation of FET geometries.
The node names are indeed based on DRAM half pitch, but CPUs haven't been made with the same process as DRAM since pitch was measured microns (eg. 0.13u).
The reality is that marketing CDs, including 32 and 22nm are only achieved through multiple patterning, and that won't change unless the industry adopts EUV or moves to a maskless process, neither of which is an economical proposition, given that the current best lightsource for EUV is a Tin vapor excimer laser, with a less than 1% dose/total energy efficiency. Part of the problem is EUV zoneplate mirrors have poor reflectivity compared to mirrors used with ArF excimer lightsources, and part of the problem is the low efficiency of the Tin vapor lasers. My prediction is that Stanford's solidstate FEL will be commercialised and used in place of excimer lasers. I'd like to see maskless litho (using DMR or DLP scanning) become a commercial reality, it will really change the market, as single dies will be manufacturable for nearly equivalent cost to volume production, but I wouldn't count on that happening too soon.
-puddingpimp
http://electronicdesign.com/digital-ics/tiny-transistors-giant-molecules-moore-s-law-crashes-laws-physics
Give this a read.
Moore's law extrapolations are hitting the limitations of physics.
As for shrinking transistors?
Pretty meaningless, silicon hit the limitations of the interconnects a while back.
Parasitic capacitance has been the brick wall that people can not get past.
www.effectiveelectrons.com "chips that work" Analog, RF, Mixed Signal
If you've ever actually had to do precision pacing and measured it out, you'd know why a pace is 2 steps. It equalizes the difference between left and right. 1% accuracy in pace length over a moderately long distance (50-500 m) isn't unusual.
As somebody who works in Lithography, I can let you know that they have not been using visible light for a long time. All fine resolution lithography is designed around as close to a monochromatic light source as possible. Having a significant spread in the light spectrum was just not consistent to do much below the 1.0 um feature size. This is because of the diffraction spread is very dependent on the wavelength and the fact that the photons have different energies thus reacting differently (or not at all) in the photoresist on the wafer.
Thus broadband lithography gave way to g-line (465nm visible blue) which gave way to i-line (365nm Ultra-Violet); Next was deep UV (248nm), Now 193nm is still used in state of the art systems today with lots of tricks such as immersion (where the light goes into water before it hits the wafer to increase the NA of the system) and double patterning (splitting up the image into multiple images that are combined in the etch processes after). Extreme UV is 13.5nm light is the next step but it is a very difficult light source to work with and the systems outrageous sums of money even for this industry.
What you are completely correct about is the importance of connectivity. I went from working in a dying 1xx nm CMOS fab this year to a thriving 1.0+um fab that makes wireless components. The lithographic part of the process (and just cramming more and more transistors on a die) is not the key value to our customers; its the exotic materials that we use to target more and more bands of wireless connectivity. I expect there will always be a demand in the market for more and faster transistors for pure computation. Its unfortunately no longer where the market growth is; thus the ROI on developing these technologies is looking more and more risky for businesses. What I have found interesting is that just about everybody working on the high end of the industry is pretty confident that the transistors will work at the 5nm-7nm node so there is still an incentive to head in that direction for now. After that will require some radical re-thinking about the materials used in computational machines.
That's also assuming they have the same Rent's exponent.
FinFET - by exerting control over the channel from three sides instead of just one - allows an improved tradeoff between leakage and speed. It's not just a leakage improvement.
Contribute to civilization: ari.aynrand.org/donate
The ultimate limit is the placement of individual atoms. It's already been done, but the process is agonizingly slow.
Contribute to civilization: ari.aynrand.org/donate
If you were doing more than thinking in tiresome categories you might have called it a self-fulfilling projection which is pretty much exactly what it became.
To refine this even more precisely, it's an ex post self-fulfilling projection, where "ex post" modifies "self-fulfilling".
But wait, there's more! It's a virtuous ex post self-fulfilling projection, where "virtuous" modifies "self-fulfilling projection".
We're now deep into The Remains of the Day. I might even call it a pink leather virtuous ex post self-fulfilling projection. Sailed through menopause without a hiccup—to everyone's great surprise—but even lathering on a hair-net bale of Grecian Formula teaser treats the glory days are well behind us.
It seems to me that the next thing to really boost computer performance is optical interconnect.
With optical interconnect, parasitic capacitance and RC delays are just gone, and associated power consumption radically reduced.
I know that there are various parties working on optical interconnect and even optical transistor equivalents.
I don't mean to imply that achieving optical interconnect (or optical transistor equivalents) will be easy, I'm just saying that it has promise to remove many of the current performance limits.
--PM
Battling for the title of who has the smallest one.
I got my first Mac just over a year ago. Because of comments like yours, I was expecting quality hardware.
It is the absolute worst computer I have ever owned. And all of my other computers were built from low-cost parts on the Internet. I've had the thousand-dollar monitor die twice (luckily under warranty) and now the video chip is flaking out whenever it displays videos. My other computers would develop issues over time, but I've never before had such serious problems in so short a time after purchase. I will never buy an Apple product again for its mythical "quality".
While I'm not a fan of Steve or Apple, Apple PCs are still superior PCs. It's just not in "geeky" stuff like processor speed or 3D performance which Apple has no control over, it's in some tangibles like quality and some other things that I personally don't give a shit about ("Design", "Form Factor").
Bull. Its noone elses problem that Mac fans insist on comparing $1200 Macs with $400 Dells. Compare a Mac with a laptop in the same price bracket, and you start to realize that there actually is competition out there. Check out the Samsung Ativ 9, or the last-year's Samsung Series 7s, or the Asus Zenbook prime.
God, shut up. Literally everything you just said is complete bullshit. Lol.. "Mains powered", huh? God, what a dumbass.
Whoever modded this post up should be flogged. I don't care if the subject is Apple or anyone else, whenever an Anonymous Coward gets on a high horse and complains about somebody's product, you do not, do not mod it up. Even if they are speaking truthfully about their own experience, if the author can't be bothered to put their name behind it (and, really, since it can be a pseudonymous throwaway account, how hard it that?), it is equivalent to bullshit, FUD, and reverse-astroturfing.
You don't reward that kind of behavior with mod points, unless it is to mod it down so that it won't be noticed by anyone else.
Motorola kept making PowerPC chips long after IBM threw in the towel, AFAIK they are still making them as embedded processors. POWER is just massive, it doesn't make sense to keep trying to make its little brother every generation.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
No-one cares about power-usage on high-end mains-powered PCs,
Bollocks. You had me until you got here, this is complete bullshit. They are called data centers, and we very much care about their power usage. They are filled with high-end PCs connected to the mains.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Of Course it does, I myself use UPS always, except for laptops, i sleep much better since then.
I use UPS too, but sometimes I use FedEx or DHL.
Apple used PPC chips (IBM-derived) for quite a few years, and at least initially they were at least as good as Intel's counterparts, often better. Before that, they used Motorola M680x0 chips, same comparison.
Originally, Intel was hobbled by their architecture. It was quirky and had a lot of historical baggage, dating back to the 8008, the first commercially available microprocessor. The cruft took chip real estate, and made it complicated to decode instructions. That was the era of the RISC chip, whose regular instruction set required very little instruction decoding. Any manufacturer who dumped backwards compatibility was able to make something faster than the latest Intel offering (although not as inexpensively). This allowed special-purpose architectures to flourish (such as Lisp Machines).
Then the CPU became bigger and much more complicated. The instruction decoding cruft took relatively far less of the chip, and more sophisticated internal processing meant that the nice regular RISC instructions lost their advantage. The CPU was munging the instructions so much that it didn't matter if they arrived as neatly typed instructions or hastily scrawled instructions, they all went into the blender and what went to the actual execution silicon was different from what went into the CPU.
At this point, RISC lost its main advantages, and CISC at least had the potential of requiring fewer instruction bytes to accomplish something. Processor power was no longer influenced by the instruction set, and Intel had more resources to throw at chip design than pretty much anybody else. This is when Macs stopped outperforming Intel PCs of their generation. When Apple in addition had problems getting laptop PPC chips comparable to their desktop chips, Apple really had no choice and went to Intel.
Nowadays, Apple computers are like most other computers in hardware, only nicer than most. When Intel Macbooks with Boot Camp came out, PC World rated them the best Windows laptops around. Add to the general build quality excellent customer support and the ability to run Unix with a nice UI, and they are much better at some things than their Windows equivalents.
"When you have eliminated the unacceptable, whatever is left, however improbable, must be the truthiness" - Holmes
Moore's law has been superseded by Koomey's law:
the number of computations per joule of energy dissipated has been doubling approximately every 1.6 years.
Koomey's law seems to hold well.
Your measurement method is bogus. You assume the entire area consists of transistors. It does not, and it never has.
If you used your method on an older "accurately sized" process, it would disagree there as well. Once again, this is because your method assumes the die is made entirely of transistors.
If you were to compute the average transistor density of a Pentium-era or P4-era chip and and adjust modern chips, then maybe your math becomes relevant.
---
According to the latest ruleset, this post should be modded as Vorpal Flamebait +5.
Shut the fuck up you pretentious idiot. Is your name really `necro81'? Is that on your birth certificate? Is that on your drivers license? Stop making useless, hypocritical, offtopic posts and kill yourself already you retarded waste of flesh.
You just missed out on the previous generation workstation that experienced the coolant leak debacle, where your Powermac G5 would suddenly leak the coolant they were using down over the motherboard and power supply and then out the bottom of the chassis.
I have one of those liquid cooled G5s. It's been an audio recording workhorse. It's going to be retired soon, but it's still going strong. I look for leaks every so often, but I've never found evidence of one. I hear that there are faster machines available these days :-)
Meanwhile the Apple doesnt have a touchscreen, and has a crappier resolution.
its not a very balanced system.
Utter bull. Intel graphics have been able to drive 1366x768 for YEARS, and somehow the latest Intel HD 5000 driving that resolution is "balanced"? The Zenbook should have no trouble whatsoever driving 1900x1080; this is the first time ive ever seen a better screen (with touch no less) touted as a flaw.
Yes, haswell is exclusive. No, it doesnt matter terribly much compared to other stats; 90% of users-- particularly those looking for a 13" laptop-- will not care about the ~10% max speed difference that it might make.
Alright, let me rephrase in that context: an Anonymous Coward post that gripes about product quality, even if the AC is speaking truthfully, does not even rise to the level of anecdotal evidence. In that light, I view it as having zero or negative merit, and should therefore never be modded up.
if the author can't be bothered to put their name behind it (and, really, since it can be a pseudonymous throwaway account, how hard it that?)
You acknowledge that AC posts can be just as anonymous as named accounts. It seems your objection is primarily to the content of this particular post.
There is a distinction between a pseudonymous account that is used once and never again, and an Anonymous Coward; they are not equivalent. I would point to this classic Penny Arcade comic. My objection is that, by being an Anonymous Coward rather than a long-established pseudonymous account, there is no way that anyone can judge what is being said: is it FUD, a bot, personal anecdote, or someone with some authority on the matter? My objection isn't with the content of the post, but rather that it got modded up, rather than staying at 0 where, as you point out, it is less likely to be seen.
it is equivalent to bullshit, FUD, and reverse-astroturfing.
I should amend my earlier statement: it may not be equivalent, but it is indistinguishable from those things, just as it is indistinguishable from fanboi-ism or astroturfing, and doesn't add to the conversation.