Become a fan of Slashdot on Facebook

 


Forgot your password?
Close
typodupeerror
×
Science

Nanomagnets Could Replace Transistors in Microprocessors 91

Posted by Unknown Lamer from the remember-when-moore's-law-failed dept.
redwolfe7707 writes "Computers today move electrons, using lots of energy in the process. A new report out of UC Berkeley shows that doing computing with nano magnetic domains could reduce the energy consumption by a factor of a million." As usual, the factor of a million would be in the ideal case and is close to the minimum permitted by the universe.
This discussion has been archived. No new comments can be posted.

Nanomagnets Could Replace Transistors in Microprocessors More

Nanomagnets Could Replace Transistors in Microprocessors

Comments Filter:

  • Dammit (Score:5, Funny)

    by LordKaT ( 619540 ) on Wednesday July 06, 2011 @09:47AM (#36671350) Homepage Journal

    Goddamn Universe! Stop treating us like we're kids! We want over a million factor reduction in our power usage! Imma go in a corner and cry :'(

    • Re: (Score:1, Offtopic)

      by dainbug ( 678555 )
      De-Regulate!
    • In the universe's defense, that limit does not apply to reversible computing [wikipedia.org]. As long as your logic gates aren't destroying any information then there is in principle no minimum energy requirement.

      • Too bad physical reversible computing hardware isn't practical. You still need non-reversible guts outside of it to track state, and reversible hardware typically can't do things like arbitrary-length loops.

  • It's only 5 years away!
    • I'm sure the start-up who wrote up the amazing summary and want investors' money also want us to believe that traditional memory will be all but dead when this comes to market. And that this is a total game changer for the entire memory industry. "Disruptive Technology" blah blah blah...

      • Re:As usual (Score:4, Insightful)

        by gstoddart ( 321705 ) on Wednesday July 06, 2011 @09:58AM (#36671428) Homepage

        I'm sure the start-up who wrote up the amazing summary and want investors' money also want us to believe that traditional memory will be all but dead when this comes to market. And that this is a total game changer for the entire memory industry. "Disruptive Technology" blah blah blah...

        I think if you came out with technology using one million times less energy, all of that would be true, no? Hell, even a factor of 100 or 1000 I should think would be a rather huge gain.

        Bring on the 5v supercomputer!

        • You would need very thick layers and cables of copper at 5v, but it's feasible. I guess you mean 5VA, that would be a real achievement ;)

          • You would need very thick layers and cables of copper at 5v, but it's feasible. I guess you mean 5VA, that would be a real achievement ;)

            Yeah ... this is the part where I admit I know exceedingly little about the physical aspects of electronics other than my high school physics, which was over two decades ago. :-P

            So ... if you say so, then that must have been what I meant. ;-)

        • Re: (Score:3)

          by PhxBlue ( 562201 )

          Bring on the 5v supercomputer!

          Pah. Wake me when they get the requirement down to 1.1 volts. I'm not going to be happy until my supercomputer can fit into a potato!

      • It's called venture capital for a reason.

        venture (noun) : an undertaking involving chance, risk, or danger; especially : a speculative business enterprise

    • It's only 5 years away!

      So we should have it 5 years before fusion power!

    • and as usual, another buzzword-using ad from one of "The Big Universities". Being from one, though, apparently allows for the publication of back-of-the-envelope calculations papers, while the rest have to submit and correct ad infinitum.

      It is like the US Patent Office and Amazon: if you work with amazon and file a patent it gets processed immediately no matter what- else, go queue up with the rest of the peasants.

    • Of course we all know that computing technology is static and changes only at geologic time scales. Don't hold your breath waiting for anything new.

      • Anybody promissing to go from idea to widespread maket in less than 14 years in computing technology (by that I mean hardware manufacturing) is a liar.

  • A problem... (Score:4, Funny)

    by ijakings ( 982830 ) on Wednesday July 06, 2011 @09:48AM (#36671358)
    Isnt it going to be more difficult to create a microprocessor with magnets? I mean nobody knows how they work...
  • This is an interesting development. If it works out, great. If it does not work out, they learn from the failure. All technology must by its nature evolve or become outmoded and worthless. As we know Computers are the poster child for technology evolution. Many more attempts to approve computers, faster, more memory, faster disks, etc have failed then succeeded. To me it does not matter if it works, IT would be nice if it did, but it is not earth shaking if it fails.

  • Fucking Nanomagnets. (Score:2, Insightful)

    by Seumas ( 6865 )

    How do they work?

  • Flakes of the future: "hold your arthritic wrist up against the magnetic CPU while running this program". "Here is a bracelet consisting of 16 broken, yet magnetic PIC microcontrollers (oh sorry, I forgot microcontrollers did not exist before the ardweeno)"

    That and I'm curious how the curie limit would affect those little things. The smaller they get the smaller the volume to hold heat, although the surface area to volume ratio improves as they shrink...

  • devices become more susceptible to random fluctuations from thermal effects, stray electromagnetic fields and other kinds of noise

    Yes, I'm imagining a big sticker on my new magnetronic iphone, depicting a big horseshoe magnet with a diagonal line through it. Airport security suddenly has to have a separate X-ray-less inspection line for portable magnetronics.

    At the moment, electrical currents are used to generate a magnetic field to erase or flip the polarity of nanomagnets, which dissipates a lot of energy. Ideally, new materials will make electrical currents unnecessary, except perhaps for relaying information from one chip to another

    Nanomagnets are hardly impressive, everything is being made "nano" these days.
    Efficient and reliable CONTROL of magnetic fields (e.g without moving electrons in an electromagnet) seems to be the critical missing piece to this puzzle.

    • Nanomagnets are hardly impressive, everything is being made "nano" these days. Efficient and reliable CONTROL of magnetic fields (e.g without moving electrons in an electromagnet) seems to be the critical missing piece to this puzzle.

      Exactly! Unless your back-of-the-envelope calculations are affiliated with one of MIT, Berkeley or Yale, in which case they are immediately eligible for publication.

  • Ok , data could be stored as a magnetic alignment, but how do you get that information from one side of the chip to the other when you need to actually use it? You can't transmit magnetism down a wire and if you use electrons then you're still going to have a large amount of wasted energy.

    • I think they were going for the calculate then transmit idea. That means only going in/out require the electrons as opposed to the present, in/out and calculate. The overwhelming majority of electrons suffer their fate during the calculate portion. Are we going to get the napkin research results hinted at? No. But if we can achieve even an order of magnitude reduction in power costs it's still "game changing" significant.
  • I like to imagine that I stay informed on science/tech stories and it seems like everyday I read about new tech and innovations.... but they never seem to be implemented in a desirable timeline. How come the big power consumers like Google, IBM and Microsoft haven't thrown loads of cash into these types of innovations? Early investments would certainly pay off in the long run- a million fold in energy efficiancy isn't good enough I guess.
    • eh?, IBM is of course doing massive amounts of research on lower power computation devices, one of the global leaders in the field

  • I see no mention of CPU speed. I'm guessing it wouldn't be that great.

  • I know people have demonstrated magnetic domain nand and nor gates, which is theoretically sufficient to build anything. Has anyone demonstrated more complex devices? A synchronous shift register? An adder? Something that demonstrates that linking many gates together is viable?

  • Reality hacked by Anonymous. Universe found to be resting on back of large turtle. Film at 11:00.

    • OMG, There are turtles all the way down!

      • Reports have started to trickle in from outlying areas stating that "The turtle stands on its own back." The turtle, in an exclusive interview with our reporters, refused to confirm or deny this, stating that it was a "personal" matter and that his first obligation was to shield his family of child universes from embarrassment.

  • ... I could get a smartphone that will last a whole week on a charge?

Slashdot Top Deals

Beware of Programmers who carry screwdrivers. -- Leonard Brandwein

Close