Slashdot Mirror
Intel Creates 30-Nanometer Transistors
SirFlakey writes: "It appears Moore's law has been proven right yet again. According to a report in Fairfax's IT section, Intel has managed to create the world's smallest transistor(s). This, according to the article would allow them to create CPU's with 10 times (420 million) the P4's transistor count. The transistors are only 3 Atoms thick(!). They say they have come close to the limit of modern technology but also still have plenty of innovation left for the future. This annoucement comes only a few days after it released an earnings warning for this quarter."
This will be like the first discussion of that kind.. oh, no, what am I thinking. Ofcourse there WILL be Intel bashing... C'mon, the guys are trying to invent/achieve something. Give them some credits.
Congrats, Intel crew!
http://dtum.livejournal.com
But didn't many of the more impressive SNES games have an additional co-processor inside the cart? Not only can you not do that with a PSX CD-ROM, it hasn't been necessary to see the same increase in apparent power.
That said, the improvement on the PSX is mostly not from tighter game code itself (not saying this hasn't happened at all though), but from ditching the Sony C libraries and coding directly to the PlayStation hardware. You almost certainly don't wan't to be doing that in a general-purpose OS.
This is my World Wide Web of Whatever
YMMV, IANMAQP, ETC
XML causes global warming.
So Intel has a transistor which is three atoms thick. According to Moore's Law, within 18 months Intel will come out with a transistor 1.5 atoms thick. Hmm, I guess the portable atom smasher isn't very far away!
/. editors have to ruin a great advance like this one by linking it to Intel's financial troubles. What are you guys saying? Intel won't be around in 18 months to top this achievement? Or is it a "But you're still losing marketshare to AMD so Nyah!" kind of mentality? Advances in science are advances in science. Just because they were made by a company with profit in mind doesn't mean their scientific discoveries won't be shared.
But seriously, I don't see why
Steven
-- I have marked myself unwilling to moderate-- I don't have other accounts to artificially inflate the karma of
--
A mind is a terrible thing to taste.
"A mind is a terrible thing to taste."
Keep in mind this article was probably aimed at the same people who buy iMacs. These people need real world comparisons to even begin to comprehend what's being said. Quoting nanometres and nanoseconds isn't going to help...
damned .. i knew I should have hit preview one more time .. ignore the bold on the 2nd paragraph please.. =\
--
Jon - TheSpork
I think it would be great if you created a major computing breakthrough! Wait, you weren't talking about you doing it were you.
This Wiki Feeds You TV and Anime - vidwiki.org
I'm entirely sure you realize this... but I'm going to state it anyway just for poops and giggles.
... He took it as a challenge and made it boot quicker. I wish more coders in this day and age took more pride in making their stuff run faster and better... rather than just running at all.
:)
Of course CPU speed isn't responsible for bootup time.. (duh) it's the OS that takes that extra time.
CPUs become faster, programmers throw more stuff at them, assuming that everybody will have the newest CPU. Why this is the norm I don't know -- I'm a coder myself and I get a chuckle everytime out of other coders saying, "Well... who cares? They can add more RAM or upgrade".
Anytime I see bootup times discussed I can't help but think of a PBS special I saw on Apple a couple of Thanksgivings ago, where "The Woz" talked about Steve Jobs asking him too make the first apple bootup quicker. Woz was happy with the time.. which says alot to me.. but Jobs wanted it faster. Rather than say "screw that"
Offtopic.. yes... moderate down as needed
Justin Buist
I think a more interesting comparison would be Super Mario World and Donkey Kong Country and Killer Instinct for the SNES. THAT was a big change...
But PC programmers are starting to get thier shit together, look at BeOS! Now if only we can get enough software ported to it...
Just when you find something good they go and cut costs.
They were going to have to cut back somewhere, heck if they can get away with using less silicon that should boost earnings. That will improve 4th quarter earnings, if not then first quarter next year.
Heck look at cars, when they started making them good and durable, that could handle a 100MPH crash. Without being totaled. Big business goes and cuts costs, making them with plastic, and aluminum. Now cars are lucky if they survive a 5mph bump in the driveway
I can see the warning on the box now: "Overclockers beware, a 2 degree increase in tempature and these silicon atoms will fuse into a new hunk of silicon."
long live the Pentium 166(non-mmx) The only cpu I have had that will running for more than a year without a reboot. While running a print/fileserver.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Some people are alive, only because it is against the law to Kill them!
Yeah, I missed the blink squared but I intentionally used mm since obviously it wouldn't be 9.8 m per blink. :P
Did you mean "One point twenty one gigawatts!!!!"
Now get back to work on that flux capacitor.
--Joe--
Program Intellivision!
Program Intellivision!
I am disheartened by your bourgousie appeal to democracy and the government as agents for change. As any student of history can tell you, government is and always has been the tool of the rich. You cannot expect an institution founded and controlled by the rich to disturb the status quo. Only a bottom-up surge of populist fervor can rock the boat enough to make it over-turn.
Moore's 'Law' is supersition; part of a belief system no more 'correct' or useful than Christianity in the Dark Ages. It's part of an economic and social system that keeps the down-trodden supressed. Any defender of Moore's Law must be prepared to defend the entire capitalist religion. Are you prepared to do this?
Good point. So perhaps IBM can etch silicon wafers down to such lilliputian dimensions, but what about thermal instability? With 3-atom-wide transistors, I'm guessing the number of electrons needed to hold a charge in a flop ain't all that much, and the alpha radiation from nearby lead (e.g. solder) could become a big(ger) concern.
Or did they forget to mention such a device is really only reliable around absolute zero?
Who said they would still be soldering parts together with lead? What about another material?
- I don't care if they globalize against free speech. All my best free thoughts are done in my head.
Actually, quantum computing is not (just) about making transistors very small. It's a totally different way of doing the computing, as many a slashdot article has pointed out :). This Scientific American article is a good overview of the subject, as well.
Never underestimate the bandwidth of a 747 filled with CD-ROMs.
the gate oxide was 0.8nm thick
Aid for the clueless: smaller transistors put off less heat so you can run them faster. Smaller transistors can be packed more closely so you can run them faster. Smaller transistors can have more of them fitted to the same chip, allowing nifty architectures so you can run things faster.
In other words, smaller = faster.
Mod down posts with a "Free Mac Mini/iPod" sig, they're spam!
Uh ... for some reason, the following phrase seems to be in contradiction with itself ...
>They say they have come close to the limit of
>modern technology but also still have plenty of
>innovation left for the future.
I must admit that I was quite surprised by the reaction my comments garnered. I didn't mean to infer that the transitor was a bad thing or a stupid invention. I think it's great that companies are continually improving products and making processors faster and smaller. I was simply pointing out the fact that the computer industry seems to sustain itself on evolutionary advances, not revolutionary advances. I consider the invention of the transistor to be a revolutionary breakthrough, I guess I'm just suprised that, to my knowledge, the computer industry hasn't embraced a wider array of research.
--It's Pimptastic!--
That's great, but what if you want to run software that was written in the last three years, or software that will be written in the next three years. Somehow, I don't think you'd like to wait a week to compile a new kernel or compress a movie into MPEG4 format. You were probably one of the ones who thought that "640k or RAM ought to be enough for anyone"
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
well, it depends :P
Are you measuring the time between blinks, or the time it takes for you to close and open your eyes? And whose eyes are you measuring, cuz politicians never blink.
My plan is to pimp before they realize I'm a jackass. Hit 'em hard and fast.
Considering quantum effects like tunelling, how exactly would you power such a 3-atom transistor processor?
It would apparently consume only very little amounts of electricity, but considering how thin the paths would be, perhaps internal resistance would rise, making temperature rise and demanding higher voltages. Higher voltages make quantum tunneling and sheer molecular structure reconfiguration much more likely.
The result would be either generalized short circuits or destruction of paths (with formation of others), I suppose.
Of course this is the first thing Intel thinks of, but it be very interesting to know how they'd manage to pull such a feat off using real world materials and at room temperature.
Flavio
Yeah, I don't know much about physics, but that doesn't stop me from arguing =).
"A good conspiracy is an unprovable one." -Conspiracy Theory
The lattice constant (distance between the center of adjacent atoms) in silicon is 5.43 angstroms. Thus one would assuem that 30nm (300 angstroms) is actually about 55 atoms thick.
Most likely the 30nm refers to the gate length and the 3 atom reference was a 'misguided' measure of the gate dielectric thickness. The reason I say misguided is because dielectrics tend to be molecules not atoms. Although 3 molecules is thin, such thicknesses have already been reported before.
So much spin. But I guess it makes sense since IEDM (International Electron Device Meeting) is occurring soon and everyone loves to get excited about the newest small transistors.
is the name of the process being used to create the chips. From a May 2000 C/Net article on the process:
"Reducing circuit size is the cornerstone of Moore's Law, which states that the number of transistors capable of being put on a processor should double every 18 months. Shrinking circuits allows manufacturers to put more transistors onto a wafer, which in turn increases power. Unfortunately, the current technique, called DUV lithography, will likely hit its limit around 2003.
Controlling small wavelength light, however, is not easy. Current lithography machines depend on lenses to focus light. Because EUV light would be absorbed by glass, the new system will use a series of four specially coated convex mirrors to capture the mask
image and reduce it. The mirrors each contain 80 separate metallic layers just 12 atoms thick.
The technology stems from work at Stanford University. The laser-light technique, meanwhile, derived from work on missile defense systems, said Dave Attwood, a professor at the University of California and a researcher on the project.
EUV machines will be able to process about 80 wafers an hour, approximately the same as current lithography machines, making the process economically feasible."
I wonder what will it cost for chipmakers to transition over to the EUV technology? Intel is huge and would obviously be more able to make a capital investment like this than competitors.
Goat sex free since 2001
And thus the great economy of the 21st century continues to thrive... If for nothing more than continued prosperity, there is value in the continued upgrading of sw/hw. However, for true productivity, most software (hello M$) is bloated, bug-ridden and have "features" above and beyond the ordinary user's needs. So, is anyone coding their web pages with Word? Betcha there are plenty who use Notepad, though...
--- Dog in, sausage out -mk
The really funny part is how many Linx running meatheads think they can second guess the most successful processor company ever.
The history of the rise of open source software, is really the story of the rise of Intel.
People try to pretend like it isn't true, but Torvalds didn't write for the 68K or the PowerPC, he wrote for the x386.
Maybe this could be used to make existing processors smaller.
:)
Maybe make wearable computers faster.
On the other hand, if they still output the same heat, but in just a smaller area, we can expect to also see flaming wearable computer operators.
How would you know if you were ripped off if you ordered a shipment of 2 million? ;)
-=-
"Everything you know is wrong. (And stupid.)"
Moderation Totals: Wrong=2, Stupid=3, Total=5.
Ah, i nver thought of that. Thats probably explains why the KI cartrage got so damn hot on my SNES...
Say that again, this time try to make some sense.
Ogg vorbis was named after ME, dangit
That the 3 atom transistor was 20 miles long and shaped like a middle finger.
Cha-Ching! ba-dump
Ok, ok, I'll stop..
Intel's gotten badly burned a couple of times lately trying to lead the market places it didn't want to go. Why on earth do you think we'd try to lead it away from "x86" any faster than the Itanium line can carry it?
Um, yes, it could. It can. It's called either ACPI S4 (Suspend to disk) (when it works) or "Bloody F*#%@#@ C{+#" (when it doesn't)...
The illegal we do immediately. The unconstitutional takes a little longer.
--Henry Kissinger
Too bad gate count doesn't necessarily = correctness or robustness of design. In fact, it's pretty much inevitable that higher complexity (in this case, by one or more magnitudes) will translate into buggier designs, unless Intel can find the time to check all circa (infinity)! operational permutations.
Is that so. hmm. I'll have to keep my eye out for that.
Whaa? "the real problem in the cache world is miss rate and branch prediction, not capacity" indeed. How do you suppose things get kicked out of cache? Little elves? While conflicts are still a problem (particularly if you don't do I-cache optimizations and/or have low associativity), increasing the capacity of the cache makes a lots of problems go away. There's a reason Intel gets to charge more for the Xeon, you know, and it's not just the groovy name.
The factual part:
To make a transistor like that work, it would have to have incredibly low resistance, gold anyone? Actually, gold would be ideal, as it can easily be made into a 3 atom thick surface. However, it could not be done with today's primitive photographic procedures. One solution would be to use a stream of electrons to shape the chip, but this is all smoke and mirrors for now.
The funny [offtopic] part:
New CueCat 2000b! Now with more features and the power of a new 1THZ (1,000,000 MHZ) processor thanks to Intel technology! Scan barcodes like never before - over 50% are scanned correctly! The most ultra-secure encryption technology protects your private information from everyone but us!
Later that day, hackers get ahold of CueCat 2000b.
Hacker: So the're still using Base64+XOR?
Hacker #2: Yep
The other funny [offtopic] part
New I-Opener 3000a! Now with the power of a 1THZ (1,000,000 MHZ) processor thanks to new Intel technology! New ultra-high tech security measures dependent on the Intel processer make the I-Openter 3000a Unhackable! This product is rock solid, with Iron Clad Security (tm)!
Later that day, hackers get ahold of the I-Opener 3000a
Hacker: Goop on the BIOS again?
Hacker #2: Yep.
And who called it a law? That is so egoist and quite frankly weird.
I wish more coders in this day and age took more pride in making their stuff run faster and better... rather than just running at all.
some do: they're called console programmers.
why do you think a 33mhz playstation went from the original mortal kombat to gran-tourismo 2 without ANY change in hardware? tighter code. the fuckwits up at redmond could learn a thing or two from these folks.
FluX
After 16 years, MTV has finally completed its deevolution into the shiny things network
"It is seldom that liberty of any kind is lost all at once." -David Hume
Intel has been using the same basic archetecture for the past 20 years.
The question that must be posed after bitching about Intel's dogged adherence to the x86 architecture is how will you get the world to change from x86 when we are already heading towards the dream of one billion connected devices, all using x86? If we suddenly decide to change to a completely new way of processing then we are going to render all of these one billion connected devices entirely obsolete - and you thought you had enough trouble keeping up with clock speed changes!
It's the same problem with the oil industry. There are too many people who have invested too much time, people and money into petroleum fuel for it to be chucked away at a moment's notice. That's the reason we're not driving Hydrogen-fuelled fuel cell cars now. So it obviously seems that if Intel won't make the switch to the next level (whatever that is) then we're going to be using the same old shit for the next 20 years!
Self Bias Resistor
Computer: A device that multiplies a user's ability to make mistakes.
----------
When the pin is pulled, Mr. Grenade is no longer our friend.
If they say they'll have them in devices within five years, they probably have a method. Besides if you actually read the article you'll see that it doesn't mention being three atoms thick, but 30nm, which is realistic, considering today's technologies are in the 130-180nm range. So 30nm is beyond the uv range, so they've devised a new etching technology. I believe the uv limit was based on the lattice spacing in what they use as a focusing lense. At any rate, the possibilities are staggering (real time ray tracing!!!), go intel, time to invest!!
Mike
Intel transfer the difficult from Hadware to software, for get more power, programmer need more technology. -- chinaitn
Indeed, let us hope that the transistor count will help the P4s future processors to not suck as much as they have been lately, I sincerely hope that they won't lose their markey completely to AMD. Competition = good.
Yeah. It took me one calculation to determine that they will need quite a few of those A Clockwork Orange eye- holder opener thingies for this to happen. That would require that of the 6 billion people on Earth, 400 million people, or 15% of the population, to buy one of these in .2 seconds. I'm not even sure if there are 800 million people worldwide with a desktop or laptop computer.
I assume they use some derivation of nano self construction. Im not exactly sure what all this entails, but it is used to create Carbon Nanotubes, and has been referenced in many other nano-scale refrences.
It expects to sell 400 million-transistor processors able to do 400 million calculations in the time it takes to blink.
Thanks for telling us the calculcations per blink, that's a real useful measurement system.
Somebody has to tell them to stop before they try to go to 1/3 of an atom. Otherwise there won't be anyone left alive to talk about it.
-
This post was compiled with `% gec -O`. email me if you need the sources
While this is cool in the fact that it will make Intel's chips cooler which means more energy efficient which translates to can be made faster, I question how this is a major breakthrough in the technology. Intel has been using the same basic archetecture for the past 20 years. I will admit that they have made many developments in size and relitive speed, but all they seem to be doing is making the chip smaller and slapping a larger heatsink on. I'll think this is much more intresting when they develop a transistor that is less than 0.05 nm.
(*disclamer* - my spelling may suck, but take the time to look past mere gramatical errors)
Macx
That's Slashdot science reporting for you. 3 atoms is completely wrong. 30 nanometers is on the order of a hundred atoms.
This advance will allow the Pentium 5 to have an all-new 700-stage pipeline to give the architecture room to be clocked up to 10 GHz. Unfortunately, due to the length of the pipeline, a branch misprediction will cause a stall lasting approximately 20 seconds. To avoid this, Intel will dedicate 300 million transistors on the chip to the world's most advanced prediction unit...
-----
Let's see yours then, coward.
The device isn't composed of 3 atoms.
It's merely 3 atoms THICK.
It could be 200 million miles long. They don't specify.
Have you seen the reprint of K&R 2? It's the same number of pages as the copy I bought several years ago, but they used such a thick stock that the book is actually thicker than O'Reilly's "C++: The Core Language". It's nearly 3x as thick as my original copy. Absurd.
(Ok, so maybe the pages aren't 3mm thick, but still...)
Something tells me the author did the calculation for 3nm, not 30nm...
--Joe--
Program Intellivision!
Program Intellivision!
Not quite... I've had this 233 at 300 for about 3 years continuously.
--
Soma: because a gramme is better than a damn.
Is translating verbal commands in real time to another language really the killer app we've all been waiting for?
Truthfully? You betcha. If there really were such a thing as three wishes, selfish wish number one would be for me to effortlessly and fluently speak all languages.
Second? For me to be able to create spontaneous multiple orgasms, remotely or up close, in any person, just by willing it.
Third? Empathy for all of the humans whom you suddenly understood from wish #1, or induced orgasm in from wish #2.
Seriously, I can't think of many greater gifts to the world than the elimination of all language barriers.
Neopets - the best free game on the Int
Getting somewhere, we just arent there yet. Anyways, I remeber hearing somewhere that the reason brain cells are so much better than transistors is that they have...many, I think it was 26 states as opposed to two. Wouldnt that be interesting if we could make a computer like that.
"Cornflakes are not the innocent critters they seem"- Sterling Morrison
Is this all in aid of bringing us yet another implementation of that wonderful, anti-orthogonal 16/32-bit dinosaur instruction set? If they spent half as much effort on Merced, or better yet, revving a really good design like PA-RISC, we might get somewhere.
PA-8200 -> 4 flops per clock sustained
PA-8700 -> 8 flops per clock sustained
Intel P4 on dual RAMBUS -> 0.14 flops per clock sustained
If / is read only, how would the system change the contents of /etc/mtab when a drive is mounted/umounted.
And remember, today is the first day of the rest of your life.
Considering that as far as our knowledge of the sub-atomics are concerned we are nearly ignorant. It is perceivable that the depth by which we are able to engineer structures is only now beginning to scratch the surface. Perhaps splitting an atom IS the next step?
Aren't the building a 50 mile long particle accelerator so they can smash these things apart and learn more about what they are?
The List of Grievances with Slashdot.
Please look up the difference between "atom" and "molecule", will you?
The illegal we do immediately. The unconstitutional takes a little longer.
--Henry Kissinger
use be and your machine will boot faster than your post will self-destruct.
--
+&x
so you find it ok that microsoft has put out an OS (windows me) that won't run on a computer that was considered state-of-the-art just about 3-4 years ago.
FluX
After 16 years, MTV has finally completed its deevolution into the shiny things network
"It is seldom that liberty of any kind is lost all at once." -David Hume
Will I be cursed as humor-impaired if I provide the blink constant: 0.33 seconds for the average human blink?
Thought so.
I am glad we have defined a new, more precise set of measurements for transistor properties and processor speed. It will, for one, make college exams much easier: Problem 1) Given 420 million transistors packed to the thickness of paper, how many operations can be made per blink?
-- Devin
-- Devin
Vivitur ingenio, caetera mortis erunt.
I mean, with all this CPU power we have, all the OS developers are starting to forget about performance improvements... They say: Why would i try to save 1 CPU clock, if most modern CPU execute more than 400 milloin of them, per second?
Yes it does, but that doesn't mean it can't happen on the scale of one or two atoms. Consider: water only converts to a gas at 100 Celcius, yet it evaporates. How? The water at the surface attains a kinetic energy sufficient to transform it into a gas. Similarly, freak quantum interactions will cause the occasional atom to convert to a different element. It doesn't take that much energy, since we're not talking about all that much mass here.
Statistically, a lump of lead could undergo spontaneous fission, but the odds of it happening are on par with the chances of every molecule of air in the room suddenly evacuating itself, leaving you gasping for breath. On a long enough time scale, everything in the universe does this; eventually it all turns into iron (and then neutronium, and then black holes, and then evaporates, but let's not quibble).
--
Dyolf Knip
"but also still have plenty of innovation left for the future."
whew! and i was afraid we were going to run out of innovation soon. thank goodness they let us know that they've stocked up.
eudas
Blessed is he who expects the worst, for he shall not be disappointed.
How about multiple CPU's on the same chip, or putting RAM on the CPU die?
I dont. I dont need a transparent "Eterm", or the-bloat-that-is-Gnome. Most people have no legal use for MPEG4 compression. The software that works today will be all I need even when I'm running it on a Pentium 12 (or hopefully, a non-chipzilla chip) in 20 years.
And I'll have you know that my precious P2/266 compiles a kernel in under 10 minutes.
400 million transistors on the same size chip that we know today will probably have the same problem but to a larger degree: interference. That many transistors so close together and so small means interference is going to be a mean factor. 3 atom thick transistors side by side just lowers the hurdle from current designs... I agree, doing a chip design to calculate not only the equations for the transistors but also the forces and fields of them all would be a bitch and a half. Personally, I'm against Moore's law.
Sorry, but having a transistor's atoms split would be nuclear fission. It takes a bloody large amount of heat to reach temperatures capable of such a reaction. Overclocking your pentium 3ghz to 4ghz isn't going to do it.
"A good conspiracy is an unprovable one." -Conspiracy Theory
one thing that amazes me even more is how much effort its going to take to actually design a chip that uses 400 mil transistors!
Not really. Any idiot can make a circuit that adds two 1-bit numbers. Any idiot can also string 128 of 1-bit adders together to make a 128-bit adder. That's how damn near *all* logic circuits are designed. Wash, rinse, and repeat. No big deal.
Just like silicon replaced by diamond - it's probably going to have a significantly (but not radically) different manufacturing process. I think that each process, such as diamond boards, or atomic transistors, will require a revolution in a particular technology, but since these benefits in process will all happen at different times I strongly believe the technology as a whole evolves. Thus, each little revolution in a particular piece of production results in the evolution of the technology as a whole.
It's a really cool idea. There is indeed a nonzero probability that all the molecules of air in a room will hit each other in just the right way so that they all leave the room through an open window. Imagine a game of pool with a few quintillion balls on a table the size of Russia, all of which are sunk on the first shot. Similar idea for sponteneous nuclear shifts.
The world of quantum mechanics is governed by laws of probabilities. There is very little that is outright forbidden, some events are just so unlikely to occur that they never do.
--
Dyolf Knip
Sigh. I don't know why I even bother to feed this troll.
/.ers. Please do us a favour -- drag yourself away from your computer, take Signal 11 with you*, and keep your evangelistical PC idealism to yourself, and let us wallow in our own X-Files induced paranoia.
They built the first transistor
Bell Labs, not Intel. Intel came up with the idea of putting tons of them in a tiny little package with little wires sticking out the side. Now they make them smaller.
Try this: build a working version of the Intel 4004 with vacuum tubes. Or worse, with the first solid-state transistor from Bell Labs, which was pretty damn huge.
Intel has always relied on the merits of its products.
Pentium floating-point bug. The F00F hack. PIII serial number. Yup, there's some pretty good things coming out there, uh-huh-uh-huh.
I know that's not fair -- the MultiBus architecture was pretty decent for its time, when there wasn't much more than VME.
Don't just spout vitriolic words from the sidelines or spin your elaborate webs of conspiracy. Your path lacks moral conviction, and I urge you to reconsider.
Jeezus. Now you're spouting out against the culture and habits of ~90% of
And no, I will not reconsider.
Thank you, and you may now mod me down.
*-Whether AM & SXI are the same is left as an exercise to the trolls.
Thus sprach DrQu+xum, SID=218745.
DrQu+xum: Proof that the lameness filter doesn't work.
Changing states is entirely different from breaking atomic bonds. When water evaporates, it's still good ol' H20, just with a bit more energy. But with nuclear fission, the actual subatomic particles are being separated, causing an nuclear reaction.
I have no idea what your point about the air evacuating itself is. There is no probability, however small you think it may be, that the air from a room will somehow disappear.
"A good conspiracy is an unprovable one." -Conspiracy Theory
This is a vaporware announcement intended to puff up their deflating stock. Simple as that.
Given Intel & Microsoft's Hegemony/Duopoly (read: making an ass load off a supply of cheap x86 hardware and a barge of ancient legacy hardware and software) I shant think we'll be seeing any revolutions anytime soon.
Wait until there's no more money left for Intel or Microsoft to make (flattening market)...then they'll try to foist something new on us and probably call it a revolution.
Wasn't Win95 supposed to be a revolution?
Nah. SNES is a great machine, but DKC and Killer Instinct had everything to gain from 32Mega(Bit!!) carts. Correct me if i'm wrong, but these carts held 4mb of data, and those were the biggest ones, and if you gzip your (illegal ;) rom images, they compress even tighter, because the snes couldn't handle decompressing all that data. the "3d" look of those games simply came from an SGI workstation. The most truly impressive game for that system, in terms of graphics was Mario World 2, with all the scaling and whatnot. Very different art style..
DKC and MW2, weighed against each other.. well, dkc has a barrel of leet, and MW2 has an entire metric fuckton of leet, to put it one way..
According to a post some time ago IBM acheived 10 nanometers as described in a pervious post. If intel claims their 30nm is smaller then IBM's 10nm they are smokeing something. http://slashdot.org/articles/00/08/12/1520241.shtm l
Aboslutely. Zaphod invended it, you know. :)
-CoG
"And with HIS stripes we are healed"
-CoG
"And with HIS stripes we are healed"
Handel's "Messiah"
Every tool reaches a level of development after which no further development is necessary or usefull. A framing hammer made today is essentially the same as a framing hammer made twenty years ago because there's no useful improvement to be made. A head machined to a nanometer's accuracy or a handle make of some wacky wundermaterial would not make my hammer any more useful to me.
Instead, progress goes into a different kind of tool. My wood and metal hammer is fine for my occasional homeowner projects, but someone with more carpentry ambition would also have a high-tech nailgun.
Same with computers. There comes a point where the typical consumer just doesn't need any more power. That's why you can still find new P-90 systems being sold - for a personal net access/word processing box, that's enough. There are many people who are no more interested in playing Quake III or doing video editing on their PC than I am in building an addition to my house as a DIY project.
Tom Swiss | the infamous tms | http://www.infamous.net/
Tom Swiss | the infamous tms | my blog
You cannot wash away blood with blood
yes
"Any connection between your reality and mine is purely coincidental." -Slashdot
"unless you go to a quantum computer which, instead of miscalculating based on the above, actually uses these principles in its functionality"
:)
...or nanomechanical computers, which dont have those probs.
Derek
You'll have to wait. He has to graduate from the 8th Grade first. Besides, he's one of the chief architects at AMD right now. You know, the guy who dumpster dives out behind Intel...
You can create an oil slick a single atom thick in your bathtub.
Dilute a thimble of baby oil with an entire bottle of isopropyl alcohol and add a few drops of food coloring to change the color (do not drink).
Fill an eye dropper and place a single drop into the bathtub. The slick will expand until it.
You can actually measure the thickness of the atom if you measure the two demensions of its area.. I'm not sure how exactly, now, but I remember doing this in my high school chemistry class.
I don't know how this applies to transistor creation, but perhaps it employs a similar tactic?
The List of Grievances with Slashdot.
not only that, but other than the fact that speed may be faster due only to the small size, how exactly does a talking computer apply to a 3 atom thick transistor? 2-way tv wristwatches? oh wait, i think they already almost have those.
Cogito Eggo Sum, I think therefore I'm a waffle
That's where IA-64 comes in. It's only because of Microsoft that we rely so much on reverse compatibility.
If the amount of heat generated is anything close to current processors, having a chip that densely populated would likely self destruct. As it is my Ahtlon 900 runs rather toasty, even with a giant copper heatsink/fan assembly.
Maybe those guys playing with liquid nitrogen as a coolant are on to something...
"They do not preach that their god will rouse them, a little before the Nuts work loose." Kipling, 'The Sons of Martha'
The inverse proportion even runs to metaphors. I remember an ad or article or something a few years ago about how this speed-demon new CPU stole the poor engineer's coffee break -- well, now he'll get it back while the damn thing reboots. Maybe with a vacation thrown in for lagniappe.
Brackets contain world's first nanosig, highly magnified:[.]
You are aware of how a transistor actually works, right? It's pretty simple... just layered silicon, simply put. Not much room to do anything except make it smaller.
However, a smaller transistor means you can cram more into something the size of your standard issue ppga chip. Mo' transistors + mo' powah - mo' size = mo' speed - mo' heat. And that is, my friend, a good thing.
-CoG
"And with HIS stripes we are healed"
-CoG
"And with HIS stripes we are healed"
Handel's "Messiah"
This means nothing but we will reach the physical limit for transistor size sooner than predicted, once we reach that barrier, the only way to increase performance would be to have larger/more CPUs. Shouldn't put more effort into inventing a quantum computer?
is it just me, or would the reductions in size follow the new innovation in fabrication technology? also, if they can now make an transistor that's only 3 atoms, what size would they be after they figure out a new way to make them.
Cogito Eggo Sum, I think therefore I'm a waffle
"Ma'am, you really shouldn't use this magnet to hold up pictures on your computer case. We found about 14 billion transistors stuck to it."
...and the new ultra-handy "What do I look like, some kinda atomic physicist?"
Or the brand new "Yeah, we had to charge you to swap out the bad electrons."
------
Let me give you the lowdown
My understanding of semiconductor design is a little shaky, but don't these devices work by localizing a charge somewhere inside them? If that's the case and the device is only the size of 3 atoms, won't it be extremely difficult to localize a charge into that space. Even geting the thing to hold a single electron seems unlikely becuase of the couloumb repulsion, spin orbit coupling, etc that souch a small device would have to overcome.
Also, if they are working with conventional processes, how will they deal with the diffraction and quantum effects of shooting electrons or photons through the mask which they use to create the chips? I'll be very interested to see the details that the article said would be realeased tomorrow, because this promises to be extraordinarily revolutionary physics if they have indeed succeeded in producing transistors this small.
> The transistors are only 3 Atoms thick
:-)
Of course its leads still had to be spaced 0.1" apart for breadboarding, but damn are they thin!
They'd just skip trying to go smaller and move directly into phase 15: Creating talking llamas whose entire cell structure is a computer.
Kinda gives whole new meaning to GIGO and WYSIWYG, eh?
------
Let me give you the lowdown
Hey, look! You're condesending! You must be Smart(tm)!
Lay off it, buddy. Nobody's impressed by your willingness to wait 5 minutes just to start your mail daemon. If you honestly think that anybody thinks you're cool because you avoid all the "new fangled computer upgrade thingies", you're fooling yourself.
Staying primitive is not a good thing. You're essentially saying, "ug. progress bad. me slow, me good!", just significantly less funny than oog was when he spoke in the same manner.
-CoG
"And with HIS stripes we are healed"
-CoG
"And with HIS stripes we are healed"
Handel's "Messiah"
No, the killer app everyone's been waiting for is a really great, free, but hard to operate, UNIX clone. Why, just the other day my grandmother was asking me for that. You'd think she'd want a computer that could translate commands like "send email" and "open Word", but apparently everyone wants to waste their time typing commands into a shell or picking out skins for their GUI. What does all this have to do with transistor thickness? Absolutely nothing!
For example:
There are, of course, other technologies other than those listed above. I'm just pointing out that it's not just that the transistors have been getting smaller (as you are too, this post would probably be better off attached to your parent)
As a slight side track, it shouldn't be forgotten that not everybody uses tiny transistors. I'm currently characterising the operation of a CMOS op-amp which uses a 2um process. One of the output stage transistors has a gate width of around 950um. Think about it, that's ~1mm.
oojah
Do you have any better hostages?
In any case, what is wrong with using a smaller process? If it makes the chip faster/cooler, I don't see any reason to complain.
--
Moore's Law isn't a fiction, although calling it a law is putting it a bit strongly. It's really just a guideline for what to expect out of the computer industry.
If you want to look for conspiricy, though, consider that Gordon Moore (who, of course, stated Moore's Law) was a major executive of Intel for many years, and CEO for part of that. I doubt there was anyone else in the world who had as much control over the rate of advance in chip technology as he did.
This one as some info on the physics in such small scale devices.
--
Jon - TheSpork
Yeah that's a great fucking philosophy. Who cares about innovation, progress, science? How about we just screw humanity's technological progress and use vi on 266's? And I'm glad you have such knowledge of future accomplishments that you know you'll never need to upgrade. Sure, running at 266 isn't that bad, but you're only at about 50% of the average computational power of the consumer. Next year you'll be at 25%, then 12.5%, then 6.75%, and four years later you'll be at about 3.5% of the average user's computational power. In other words, it will take you 27 times longer than the average user, whose computer will be targeted by software developers. Not only will it be impractical not to upgrade, it makes no economic sense, unless your time is worth significantly less than the minimum wage. On one hand you will be able to purchase a pentium 5 2ghz for $100, and on the other you can spend six hours recompiling your kernel. Computers will have become increasingly relied upon in everyday use, and I'd bet that 6 hours is worth more than a hundred bucks.
"A good conspiracy is an unprovable one." -Conspiracy Theory
Is this news for nerds? or is this "hey we want to be market analysts"?
I ate my sig.
> how much effort its going to take to actually
> design a chip that uses 400 mil transistors!
If you tried to do a custom design flow for all 400 million transistors, then, yes, your design cycle would be crazy bg. But, if you used VHDL/Verilog and an ASIC-type flow with custom design for only the most speed-critical pieces, it's possible. In an HDL, one line of code (like a = b*c + a) can translate into many thousands of gates.
And the next time my disk fails and my backups are inaccessible, can I count on you to remind me that Murphy's law isn't a law, either? Or the next time a usenet thread degenerates into argument-by-nazi-allusion, that Godwin's law isn't a law, either?
The law of gravity is merely a sophisticated conjecture fitted to manifest evidence. Moore's law is no different, and your need to find malicious economic motives is most disturbing. Intel is in the business of making fast processors, and if there's any way they could make a few technical changes and build a better processor, they would. That's what they've always done, and that's what they'll continue doing. Just ask AMD.
They built the first transistor (well, technically they didn't exist at the time, but it was the same brains that founded Intel), so they're the ones who've been doing it the longest. If there were a means, then they'd pursue it. Unlike companies like Microsoft which start with a few engineering innovations and coast on their legal verdicts from there, Intel has always relied on the merits of its products. They couldn't have gotten to where they are today if they hadn't.
And if you disagree with how much money Andy Grove has earned from his efforts, then do your part to vote and raise his taxes and finance a better wealth-redistribution scheme in this country. Don't just spout vitriolic words from the sidelines or spin your elaborate webs of conspiracy. Your path lacks moral conviction, and I urge you to reconsider.
-- Anne Marie
Yeah, and 400 million transistors gives a lot of room for design slop--more space for slapping together pre-designed components.
Any idiot can make a circuit that adds two 1-bit numbers. Any idiot can also string 128 of 1-bit adders together to make a 128-bit adder. That's how damn near *all* logic circuits are designed. Wash, rinse, and repeat. No big deal.
Sure, any idiot can string together 128 1-bit adders, but designing a 128 bit adder to run at that high of a clock speed takes a bit more work. It would have to use some kind of carry-lookahead logic trick to get everything it needed done in one cycle. Point being, putting together a solid, optimized component like that DOES take some serious design time--if for nothing else to but to do the math using a CAD program or espresso. And if that takes effort, getting your stuff to play nice at a high enough clock speed must take more!
I'm far and away no pro [yet] at this sort of thing, but from what I've done myself so far (just introductory digital design stuff, building components and simple clocked machines) it would take a long time to put together something this complex and do it right. Witness the P4.
-s
- - - - - - - -
Don't worry, being eaten by a crocodile is just like going to sleep in a giant blender.
Customer: Your program crashes randomly. It must be a bug.
Support: It's not a bug. It's a quantum fluctuation.
The description only mentions the thickness, not how wide or long these transistors must be. Perhaps they can build them sideways where they need to fit a lot of them together.
It seems like at that level of thickness you would have to start concerning yourselves with the crystaline structures of whatever you're using to insulate between layers of transistors.
With it only 3 atoms thick you'd think that there would be fab screwups causing bands in the transistors to narrow to an unusable level - probably happening quite frequently. Would play hell with your yield thats for certain.
I wonder, though, if they're doing work with transistor area. If a reduction to 3 atoms thick bought them another 10 years of industry life I wonder what shrinking the sides by 1/2 would do.
Ad in classifieds: Pandora's Box (no box) $5
--
A mind is a terrible thing to taste.
"A mind is a terrible thing to taste."
The x86 line was never meant to lead to heavy duty CPUs, and it has never been a heavy duty CPU. Its a consumer grade CPU, and it does well as a consumer grade CPU.
Some fluke long long ago lead us to settle on the x86 line for PCs. With all the backwards compatability in the x86 line and microsoft DOS/windows9x/windowsME, I don't think twice about running some old DOS program that last executed under a 286 processor running at 16 mhz on my current win98SE machine with a CPU a few generations later. The whole force behind the PC market is that Joe Consumer can upgrade his machine or buy a new one and not have to worry about his old software. Do you think PCs would be as cheap, and as advanced as they are if every new PC meant that Joe Consumer had to buy an entire new software package?
If you want a real cpu for heavy work, then avoid the x86 and grab a nice unix varient for your OS. Don't complain because you buy a desktop machine and discover that it was made for the mass market consumer, and not for heavy duty work.
Choose your tools carefully, and remember, 99.9% of all the people out there are so perfectly happy with the x86 architecture, they have no clue that anything else exists.
Do no try to get a decent word processor on punch cards, instead, try to punch holes on cards by yourself, then u will found that there is no word precessor.
i don't think we're quite at nanoconstructiony yet... carbon nanotubes are made in arc reactors by vaporizing graphite in a low pressure inert atmosphere. most likely, intel's tampering with a new optical lithography technique. joe
With these chips, computers will be able to translate verbal commands or conversations from one language to another in real time, or search massive and complex optical databases.
Don't you just love the examples that are used to "show off" the speed of new chips to the masses? Is translating verbal commands in real time to another language really the killer app we've all been waiting for?
You: "Cocine una cena para mí!"
Computer: "Screw you."
Never underestimate the bandwidth of a 747 filled with CD-ROMs.
Wrong.
You obviously know very little about computer engineering. There have been many, many advances besides fitting more transistors on a chip.
1. Out-of-order instruction scheduling (branch prediction, pipelining, etc.)
2. RISC design 'philosophy'
3. IC Design tools (simulations, etc.)
4. Advances in vital subsystems (memory bus, RAM, controllers, etc.)
These techniques would have all been possible without transistor size advances beyond the original 8088 level. And you could probably name a few more but I don't know that many yet.
Realistically speaking, we don't know what the future will hold for computing. People are talking about quantum devices (have interesting mathematical properties) and optical devices (good for fiber optics)
"Any connection between your reality and mine is purely coincidental." -Slashdot
the article's referring to two different dimensions- the 30 nm is the width of the transistors, the 3 atoms is the thickness. putting down a 3 atom thick layer of anything is easy with vapor deposition (an extra atom here or there probably doesn't make a difference). I'm sure nano-assemblers would be lovely, but you can put down an even coat with $200 worth of vacuum sputter coating equipment now. The real impressive aspect of what they're doing is the "30 nm wide" part.
Your allusion is quite faulty: a hammer is useful for an incredibly smaller number of things than a computer, whose processing power has unlimited applications. Sure, for word processing and net access a p90 might be sufficient, but as computing power and software development progress, more tasks can be made more efficient, quicker, and easier by shifting them to computers. Things like digital cameras and video cameras allow the consumer to take many more pictures than the analog variety, and are continually decreasing in price and increasing in features.
I'm sure that if building an addition onto your house cost half as much every year, you would do it, and then as the prices became even lower, you would upgrade again in a couple years, and again, and again. The fact of the matter is the Parkinson's Law of Data ("Data expands to fill the space available for storage") is true for everything, and follows the basic economic theory of opportunity cost.
It is more efficient in all aspects for consumers to regularly upgrade their computers than it is for them to continue to use outdated models.
Sure, right now the average family might think that they only need your computer for word processing, email, and basic web surfing, but think back to ten years ago. Word processing was almost unknown, the web was unknown, yet now my teachers require all papers to be typed. Soon all non-presentation-based homework will be sent in using secure web applications.
Society will assimilate technology infinitely, and by catering to the average user, software will mandate regular upgrades for increasingly computer-based social functioning.
"A good conspiracy is an unprovable one." -Conspiracy Theory
Sharkey Extreme also has an article about this located here.
he might be wont to say "Ich bin ein BEOWULF CLUSTER!"
--
A mind is a terrible thing to taste.
"A mind is a terrible thing to taste."
Every major advance in the last 40 years has been due to increases in clock speed and switch density. Cute tricks like caching and dual-piping or whatever they're calling it this year are flea bites on the butt of real progress. Remember what an "advance" the 486 was over the 386? The corporate boojums need things to market so they make things up when there's nothing real in the pipe, but when something real comes along it doesn't have to be marketed to you because you sure as damn hell notice it.
I mean, my relatively nonobsolte PIII is real cool, but would it really be that much cooler than a machine with 486-level architecture running at the same 450 MHz? For that matter I have to wonder how my tired old 8-bit friends would fare if one could run them at a good fraction of a GHz. Sure, you buy some extra clocks with all those extra transistors trying to second-guess look-ahead your code, but I wonder if that's the best use of all that high-speed silicon. Maybe a *cough* beowulf cluster */cough* of, say, Z80-level CPUs all fabbed on one chip and running at 1GHz could do some really interesting things by comparison.
If this thing is real then great for Intel and for us, it doesn't really matter what architecture they apply it to; and if it isn't real it won't save them when something that is does come along, not matter how good their press releases are.
Brackets contain world's first nanosig, highly magnified:[.]
I kind of hope not, since when you split an atom it violently explodes.
Though, that may be next up for Creative when they come out with the "nuclear vlolitile mega monster blaster annihilator pro plus atomic destroyer!" series of 3d cards.
-CoG
"And with HIS stripes we are healed"
-CoG
"And with HIS stripes we are healed"
Handel's "Messiah"
Gallium Arsenide.
"Be polite, be professional, but have a plan to kill everybody you meet." General James Mattis
Assuming you have an origional (unlocked) P2 (as I do), you could get a decent heatsink, change some mobo jumpers and you could be running at 300MHz in a few minutes...
--
Soma: because a gramme is better than a damn.
I was just talking with a collegue working on Bose-Einstein condensations (BEC) and I asked what some of the uses were. Due to the way BECs work statically/quantum mechanically one can create any interferance pattern within the BEC. He said that there are people working on trying to figure out ways of using this property to replace the etching processes used today to create things like computer chips by creating a interferance pattern in the form that one wants and then laying the BEC on the matterial (there is more to it than that but you know that). This would allow for manufacture of things at the 3 atom level. Of course, as someone else mentioned, 30 Nanometers is larger than 3 atoms thick. Lattice structures of silicides are roughly between .1 and .9 nm [1].
Theoretically this is possible, now whether this is practical is a whole different ball park.
[1]V.E. Borisenko: Semi-conducting Silicides (Springer, New York): pp 3-5
Disclamer - Opinion of Person
Actually, we don't use the same thing that was invented in 1947 for ICs now. There are all kinds of transistors. BJTs, IGBTs, FETs, MOSFETs. The latter being the type used in modern semiconductor technology. Forgive any errors (I've not yet taken solid state), but whereas a conventional transistor emits a collector-emitter current proportional (the gain) to the base-emitter current, a MOSFET's gate is a capacitor (in fact the capacitors used for DRAM are just MOSFETs) where the current through them is proportional to the voltage across the gate. They are much more disposed to on-off operation than operation over a linear region, because it requires minimal (gate capacitor leakage current) energy to maintain a MOSFET gate state, whereas to represent a '1' on a BJT would take a constant supply of current, irregardless of whether it had changed recently or not.
BJT = Bipolar Junction Transistor
IGBT = Insulated Gate Bipolar Transistor
FET = Field Effect Transistor
MOSFET = Metal Oxide Semiconductor Field Effect Transistor
As you can see, there have been many advances more significant than having the boys in the back room develop a better/smaller/faster/more powerful widget.
Size does matter. The major factor limiting the speed of something like a CPU is the time that it takes for an electrical signal to propagate from one point to another. During each clock cycle many signals have to move across the CPU die. You can make the CPU bigger, but then you have to run it slower to allow for the increased signal propagation time.
Given the smaller and smaller transitors, I have always wondered what the effect of ionizing radiation is on these things. Granted, your average cpu is not out in outer space someplace, but even your everyday enviroment has its share of crud running around (Xenon from granite ferinst). Are we going to have to be careful about protecting these tinie weenie gates that use using very, very few electrons, or are we going to have to build error detecting/correcting logic into the cpu itself?
Making a small transistor is meaningless. Making a huge amount of small transistors packed together which don't produce too much heat and require only a few Watts is what is necessary for this to be meaningful. I don't think it's meaningful yet. And remember, rambus.
/me imagines a building (city?) filled with 400 million tubes. Wheeee!
---
Video meliora proboque deteriora sequor - Ovidius
Rather than looking into new and innovative ways to increase a CPUs power like Sun does with their Ultrasparc line or Digital did with their Alphas or even AMD has done with their line of chips, they just try to keep shrinking the size of the transitors, pumping more of them into the CPU, and ramping up the clock speed. When are they going to learn that the x86 architecture is dead dead dead? I REALLY hope they don't screw up Merced.. err.. Itanium by keeping the prices too artificially high. I'd really like to see that technology move into consumer PCs instead of just servers. We need a stepping stone out of the x86 world while preserving the cost factor that makes x86 based systems a more palatable choice over the higher end and more expensive workstations.
Well....it states they could make a processor with ten times as many transistors.....OR....they could make the same processor ten times smaller! Why not do that?
-Just a thought....
-Julius X
-Julius X
remove "-whatkindofspamdoyoutakemefor-" from email to send
Good lord thats a lot. This should fill in nicely while molecular computing advances to the point of commercial feasibility as a technology.
:o)
However, one thing that amazes me even more is how much effort its going to take to actually design a chip that uses 400 mil transistors! I'm a computer engineering student: designing small stuff using just a few is enough for me.
I guess Intel'll be hiring soon.
-S
- - - - - - - -
Don't worry, being eaten by a crocodile is just like going to sleep in a giant blender.
Quantum computers. The ability to get millions of transistors using a few atoms.
The problem is not so much of a capacity misses as much as conflict misses... really. I have ran simulations of processors with varying cache size with suprisingsly little or no improvement beyond a 16 K instruction cache using the Trace Cache method.
It is 'how' that sequence of streams are fetch, not 'how much'. You can fetch the whole 16K block of instruction, but if the insturction tend to jump to the other end of the memory location (not within the 16K window), you just spend x amount of cycle doing crap.
If you talk of capacity, the only real sufficient capacity is the cache sufficient to load up the whole instruction segment.... that way we wouldn't have instruction miss (75% of the references!).... while we are not yet there, just work on a better branch prediction.
As I am sure some people have seen, it is possible to deposit single atoms, and react two specific atoms (leaving others untouched) using a technique that uses a scanning-tunnleing microscope. This could probably do the trick, but, again, raises very important questions as to how to mass produce the things. As a chemical engineer, I'll believe it when I see it...or my boss tries and make me design the blasted thing. Then again, as a chemical engineer, I can't see why you would want to just react two atoms at a time anyway ;-)
Don't blame Intel on that... blame the vendor of whatever OS you happen to be using... Though, from what i've heard, BeOS boots within 10 or 15 seconds and Mac OS X is supposedly going to power right on up as well, so it's doable.
Couldn't an OS take a hardware inventory and mirror its ram to disk on shutdown, then at startup, if the BIOS didn't report any changes to the hardware configuration, simply load the last memory image and forget about have to go through the entire boot process?
Actually, 640k would suit me fine.
In my processor cache.
WARNING: there is a trojan on your
30 nanometers refers to the width of the transistor gate. not the thickness/height.
From the article:
The No. 1 computer-chip maker plans release details in the US on Monday about the transistors, which are just 30 nanometers - three-millionths of a centimtere - thick.
Where are you getting your information from?
--
A mind is a terrible thing to taste.
"A mind is a terrible thing to taste."
Does not equal faster cycle time as much as faster cycle does not necessary mean faster computer (to scale with cycle time reduction).
However, more transistor means that we can afford more internal cache... but then the real problem in the cache world is miss rate and branch prediction, not capacity.
So what to do with all those unused transistors? System on chip is one of the possibility...any other suggestions?
This is not an issue for laptops, PDA's, or the physical size of the computer under you desk. What this does affect is VLSI (ULSI?) IC's. Reduced tranistor size means lower operating voltages. Which means you can scale down supply voltage, which reduces electric field strength and power dissapation in the tranisistors. This leads to boosted device density and switching speed. In short this allows VLSI designers to create faster, more complex and powerfull IC's and/or ones that require less power(this could affect your laptops and PDA's). As for powerfull and complex look at the IBM Power4 processor it contains 170 million transistors, SIA (Semiconductor Industry Association) predicted in 1999 that by 2002 microprocessors would contain 76 million transistors !!! This is all because of the incredible shrinking transistor.
;)
It's really nice that you think its time for a major computing breakthrough. Personally, I think its about time for a major transportation breakthough, something that really catapults transportation into a new era, not unlike the invention of the wheel itself
- remove the primate to mail
Are these devices composed of only a single element? If not, they must be molecules. How did they get (at most) 1 molecule to act as a transistor?
Yep, I never spell check.
More incorrect spellings can be found he
A stack of more than 100,000 of the 30-nanometer transistors, which act as switches to control the flow of electricity in a chip, would equal the thickness of a sheet of paper
--
A mind is a terrible thing to taste.
"A mind is a terrible thing to taste."
30 nanometers refers to the width of the transistor gate. not the thickness/height.
Before transistor there were tubes. Some people
still swear by them. This transistor revolution
you speak of, is really all about scalability/power
consumption. There was nothing more revolutionary
to Bardeen's baby.
Mechanical penguins love transistors. The Linux Pimp
--It's Pimptastic!--
Well, at least Intel does *some* good for the world. I heard the chip comes with a free tracking device included. :)
Moving at the speed of government.
Intel's new Marchitechture... 6 years away, if they solve all the heat issues. I dunno about the rest of the /. audiance, but my P2/266 is still plenty fast. It runs fvwm2 as fast as the day I got it, three years ago. And I predict it will run xterm just as fast in another three years.
What I can't see is how one can lay down anything 3 atoms thick (or wide) reliably (in the sense of real-world mass manufacture, not one of a time in-the-lab productions) using scaled versions of existing Fab tachnologies and without some nano-assembler type technology. Worst case you'll get 3 atoms somewhere in the middle of the wafer and maybe 5 or 0 at the edges ....
This sort of tech will come one day - but I beleive it's going have to be by revolution, not evolution ....
Spoken like someone who has never taken Quantum Mechanics 101.
XML causes global warming.
It ought to be instant on by now. Sheesh.