If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Abc.net.au)   Scientists break Moore's Law by creating single atom transistor 8 years ahead of schedule   (abc.net.au) divider line 13
    More: Cool, Moore's Law, atoms, hydrogen atoms, Nature Nanotechnology, transistors, University of New South Wales, qubits, vacuum chambers  
•       •       •

6329 clicks; posted to Geek » on 20 Feb 2012 at 12:07 AM (2 years ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



Voting Results (Smartest)
View Voting Results: Smartest and Funniest


Archived thread
2012-02-19 08:42:44 PM
3 votes:
Manufacturability is another matter, submitter.

Moore's law describes a long-term trend in the history of computing hardware whereby the number of transistors that can be placed inexpensively on an integrated circuit doubles approximately every two years.

Cool but if they can only make one, it doesn't count.
2012-02-20 12:35:37 AM
2 votes:

cman: Oh, and now intel might be able to break the 4Ghz stock speed


No it won't. Heat output of a CMOS circuit scales with frequency. Modern clock speed limits have nothing to do with feature size and everything to do with the inability of air-cooled finned heatsinks to effectively remove more than ~150 watts of heat from a one square centimeter area without sounding like jet engines or having an unacceptable temperature at the hot side, known as the Thermal Brick Wall.

And no, Joe Average is not going to install liquid cooling until it's maintainence-free for life the same way we expect every other piece of computer hardware to be.
2012-02-20 08:39:30 AM
1 votes:

cman: Oh, and now intel might be able to break the 4Ghz stock speed


Clock speed is no longer a meaningful measure for a computer's effectiveness.
2012-02-20 08:32:36 AM
1 votes:

Chiad: PerilousApricot: Chiad: gaslight: Manufacturability is another matter, submitter.

Moore's law describes a long-term trend in the history of computing hardware whereby the number of transistors that can be placed inexpensively on an integrated circuit doubles approximately every two years.

Cool but if they can only make one, it doesn't count.

Read it again, many have made a single atom transistor, they have worked it to a level of precision that is repeatable and useful for manufacturing.

Nope, placing atoms individually with the tip of a electron microscope isn't inexpensive, and certainly won't scale to the level that you could build an actual computer chip with it.

It's still a cool result, but I wouldn't exactly be popping champaign over it.


Repeatable and useful, not inexpensive, they still have to work that part out, I'll grant you. But this isn't the first single atom transistor.


As an integration engineer who works in cutting edge semiconductor research, I get a kick out of these articles. Don't get me wrong, it's really cool stuff, but I think most people don't understand the gap between a university lab result and an actual product. Did you notice the size of the sample? Today's work is done on 300mm wafers with some lookahead work started on 450mm. The reason? We scale primarily for cost, not for performance. It's not about having smaller chips to fit into smaller packaging because no consumer product is going to be much smaller than they are now because they need to be handled, but rather, we're trying to get to more chips per wafer. If each generation isn't cheaper, then it's not going to be made. So far the industry has been able to keep the pace, but it's getting much more difficult.

Also, a functioning chip has many other elements so it would be curious to see how you would integrate this process with all the passives and other devices.

Anyhow, I'm not trying to pee into anyone's Corn Flakes, but just to level-set expectations.
2012-02-20 08:04:17 AM
1 votes:

BigJake: bingethinker: Go do your homework, junior, the adults are talking here.

he's right


Not in 1982 he wasn't. In 1982, your typical choices were:

The original IBC PC with a Intel 8088 (16 bit, ~ 1/2 MHz), 16 K of RAM, one floppy drive & no hard-drive, running PC-DOS, for $2,945.

An Apple II Plus with a MOS 6052 (8 bit, ~ 1 MHz), 16 K of RAM, one floppy drive & no hard-drive, running Apple DOS, for $1,195.

At the time, Apple's computer was clearly the better choice for home users. This stayed true in 1983, with the release of the IBM PC XT and the Apple IIe.

Where things started to diverge was with the 1984 release of the Macintosh. With PCs stuck and DOS and Apple rolling out a graphical UI, Apple could charge a premium and people would pay gladly. It wasn't until 1992 that Microsoft had a workable competitor with Windows 3.1, but that still ran on DOS. You can't say that Microsoft really caught up until Windows 95, released 11 years after the Mac.
2012-02-20 07:52:44 AM
1 votes:

erik-k: cman: Oh, and now intel might be able to break the 4Ghz stock speed

No it won't. Heat output of a CMOS circuit scales with frequency. Modern clock speed limits have nothing to do with feature size and everything to do with the inability of air-cooled finned heatsinks to effectively remove more than ~150 watts of heat from a one square centimeter area without sounding like jet engines or having an unacceptable temperature at the hot side, known as the Thermal Brick Wall.

And no, Joe Average is not going to install liquid cooling until it's maintainence-free for life the same way we expect every other piece of computer hardware to be.


Liquid cooling is easy. Just drop your PC into an aquarium filled with mineral oil (new window).
2012-02-20 05:37:48 AM
1 votes:

erik-k: And no, Joe Average is not going to install liquid cooling until it's maintainence-free for life the same way we expect every other piece of computer hardware to be.


Joe Average doesn't install his own computer components. He buys a Dell and throws it out when it meets its planned obsolescence in 3-5 years and buys another Dell.

For those who do install their own components (Joe Above-Average), all-in-one coolers like the Corsair H80 are brilliant, and yes, maintenance-free (and much smaller than the giganormous air-cooled sinks you see these days).
2012-02-20 04:10:40 AM
1 votes:

Chiad: gaslight: Manufacturability is another matter, submitter.

Moore's law describes a long-term trend in the history of computing hardware whereby the number of transistors that can be placed inexpensively on an integrated circuit doubles approximately every two years.

Cool but if they can only make one, it doesn't count.

Read it again, many have made a single atom transistor, they have worked it to a level of precision that is repeatable and useful for manufacturing.


Making it precise enough to be repeatable doesn't mean it's cheap enough to be manufacturable. Totally different goals. At best, within 10 years you'd have other groups repeat the experiment.
2012-02-20 03:17:00 AM
1 votes:
You can tell the article was written with Americans in mind. It includes a link to Google maps showing what an Australia may look like.
2012-02-20 01:30:18 AM
1 votes:
Aaaaand done in the Boobies. Moore's Law is an axiom that refers to what the consumer sees, not what R&D does.
2012-02-20 01:04:27 AM
1 votes:

Chiad: gaslight: Manufacturability is another matter, submitter.

Moore's law describes a long-term trend in the history of computing hardware whereby the number of transistors that can be placed inexpensively on an integrated circuit doubles approximately every two years.

Cool but if they can only make one, it doesn't count.

Read it again, many have made a single atom transistor, they have worked it to a level of precision that is repeatable and useful for manufacturing.


Nope, placing atoms individually with the tip of a electron microscope isn't inexpensive, and certainly won't scale to the level that you could build an actual computer chip with it.

It's still a cool result, but I wouldn't exactly be popping champaign over it.
2012-02-20 12:48:56 AM
1 votes:

Electrify: So how long before we make transistors made out of a single electron? Or are we just going to start splitting atoms to squeeze more if them into the processor?


the next advancement seems to be carbon and silicon nanotubes...current two-dimensional processors are inefficient compared to three-dimensional processors.
2012-02-20 12:10:38 AM
1 votes:
Actually, this is right on time. Given 8 more years, the technology will be mature enough to create trillions of them on a single surface cheaply and quickly, and sell on the open market for use. They didn't beat the law, it feels like they're upholding it.
 
Displayed 13 of 13 comments

View Voting Results: Smartest and Funniest


This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »






Report