Slashdot Mirror

← Back to Stories (view on slashdot.org)

Study Opens Route To Ultra-Low-Power Microchips (mit.edu)

Posted by msmash on from the shape-of-things-to-come dept.

Freshly Exhumed writes: A new approach to controlling magnetism in a microchip could open the doors to memory, computing, and sensing devices that consume drastically less power than existing versions. The approach could also overcome some of the inherent physical limitations that have been slowing progress in this area until now.

Researchers at MIT and at Brookhaven National Laboratory have demonstrated that they can control the magnetic properties of a thin-film material simply by applying a small voltage. Changes in magnetic orientation made in this way remain in their new state without the need for any ongoing power, unlike today's standard memory chips, the team has found. The new finding is being reported today in the journal Nature Materials, in a paper by Geoffrey Beach, a professor of materials science and engineering and co-director of the MIT Materials Research Laboratory; graduate student Aik Jun Tan; and eight others at MIT and Brookhaven.

As silicon microchips draw closer to fundamental physical limits that could cap their ability to continue increasing their capabilities while decreasing their power consumption, researchers have been exploring a variety of new technologies that might get around these limits. One of the promising alternatives is an approach called spintronics, which makes use of a property of electrons called spin, instead of their electrical charge. Because spintronic devices can retain their magnetic properties without the need for constant power, which silicon memory chips require, they need far less power to operate. They also generate far less heat -- another major limiting factor for today's devices.

41 comments

  1. Toriod core memory anyone? by bobbied · · Score: 2

    Sounds like an improvement on core memory..

    --
    "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
    1. Re:Toriod core memory anyone? by ShanghaiBill · · Score: 2

      Sounds like an improvement on core memory..

      ... in the same way that fractional-step UV photolithography sounds like an improvement on a hammer and chisel.

    2. Re:Toriod core memory anyone? by bobbied · · Score: 1

      Sounds like an improvement on core memory..

      ... in the same way that fractional-step UV photolithography sounds like an improvement on a hammer and chisel.

      Maybe, but it's basically a smaller version of the toroid core memory we knew and loved 40 years ago (and yes, I've actually used the stuff). I remember looking at it though a microscope trying to figure out how they actually built this stuff that small?

      I'd call it roughly the same kind of advance we made from transistors to VLIS chip design.. Only without the speed increase. :)

      Power consumption being lower, not requiring power to maintain it's memory, are all great, but if reading the data takes too much longer than DRAM or Flash, it's all more interesting than actually practical. Maybe in applications where core memory once was the choice, or as a backing store for DRAM for things like RAID controllers when the power fails or controllers where it would be nice to remember small amounts of stuff over long power off cycles. Maybe in space applications, where long term temperature and radiation stable storage without having to consume power is more important than speed?

      I'm just not seeing the killer application here...

      --
      "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
  2. Will storage and memory eventually become one? by SuperKendall · · Score: 2

    Reading through the abstract something that struck me was the statement "with no degradation in magnetic properties after >2,000 cycles".

    With the increase in speed of SSD's all the time, and advances like this that don't suffer degradation, it made me wonder if at some point there would be no need for separation of RAM and SSD, if storage were fast enough you could just use as much of it as you liked for system memory.

    Looking around at some specs it seems like at this point RAM may be just 10x faster than the best SSD's around, probably less now. I'm sure there will always be even faster L1/L2 cache memory chips to speed things up, but just thinking of the system RAM we all have today - there has to be a point where the primary storage is fast enough to take on that role and gain greatly improved system memory as a result.

    --
    "There is more worth loving than we have strength to love." - Brian Jay Stanley
    1. Re:Will storage and memory eventually become one? by fahrbot-bot · · Score: 2

      Reading through the abstract something that struck me was the statement "with no degradation in magnetic properties after >2,000 cycles".

      Just don't use it *before* 2,001 cycles, 'cause, man, it's all over the place.

      --
      It must have been something you assimilated. . . .
    2. Re:Will storage and memory eventually become one? by Anonymous Coward · · Score: 0

      Probably not for a while, latency is the important benchmark for RAM, not throughput, some quick looks appear that RAM latency is typically measured in nanoseconds while SSD latency is measured in milliseconds. Some of the high performance SSDs seem to be able to manage around .2 ms, but when a lot of RAM is managing in the order of 10 ns, that's 3 orders of magnitude slower for the SSD. That's a lot of ground to make up.

    3. Re:Will storage and memory eventually become one? by shess · · Score: 2

      Reading through the abstract something that struck me was the statement "with no degradation in magnetic properties after >2,000 cycles".

      With the increase in speed of SSD's all the time, and advances like this that don't suffer degradation, it made me wonder if at some point there would be no need for separation of RAM and SSD, if storage were fast enough you could just use as much of it as you liked for system memory.

      Looking around at some specs it seems like at this point RAM may be just 10x faster than the best SSD's around, probably less now. I'm sure there will always be even faster L1/L2 cache memory chips to speed things up, but just thinking of the system RAM we all have today - there has to be a point where the primary storage is fast enough to take on that role and gain greatly improved system memory as a result.

      For a long time, the time it takes the transfer data from RAM to the CPU has been the bottleneck in performance. That is a large part of why so much of a CPU's die is devoted to cache - putting a ton of computational elements on the die is pointless if you can't keep them busy, and cache is also more straightforward to design. Switching out RAM for flash would cause a HUGE performance hit.

      Beyond that, RAM can cycle continuously with no upper limit. Due to caches, most DRAM isn't going to cycle as frequently as SRAM in the cache will, but 2000 cycles is a trivial number when you're talking about RAM, you can pass that in fractions of a second. Many people are uncomfortable using SSDs for swap, which is a primitive way of using it as RAM.

      I would be very surprised if we ever reach a point where there's some sort of unified storage and memory hardware, unless we freeze the target. Basically, anytime someone invents new technology, we generally use it to expand the parameters of the space to address new problems, rather than only using it to simplify the existing space.

    4. Re: Will storage and memory eventually become one? by Anonymous Coward · · Score: 0

      I have noticed that these research projects use very low tolerances for disruption or other measurement error. In real world applications these can often be relaxed, enabling possibly improved results. I hav never read the author previous works but it would be interesting to compare their previous lab results with the ultimate applications.

    5. Re:Will storage and memory eventually become one? by SuperKendall · · Score: 1

      That's why this new storage technique could bring the two together I was thinking - if it could reach the speed (and as someone else pointed out) the latency of RAM, and also has no cycle limit (that's what the >2000 cycles part is leading me to think might be achieved) and it can store data without power as can SSD, then perhaps you could use it for both purposes.

      Not really enough detail in the summary to suggest this particular tech could or would be the one able to do so. But the combination of properties made me think, maybe someday the two could be merged without performance loss (or with some tolerable degree of performance loss given the benefits of combined forms of computer memory).

      --
      "There is more worth loving than we have strength to love." - Brian Jay Stanley
    6. Re:Will storage and memory eventually become one? by bobbied · · Score: 1

      That's why this new storage technique could bring the two together.

      I doubt it's fast enough for even being used as disk space, at least compared to flash. Unless they have some novel way to read the magnetic fields, this kind of indirect sensing is pretty time consuming. The way you detect a magnetic field is by creating one of your own by passing a current past the cell you are trying to read. This requires you to carefully control the "read" pulse and watch for tiny variances in the rise and/or fall times. This kind of thing is slow, at least it's slower than just looking at the voltage level of a memory cell. Writing is similar, only likely even slower, because you have to actually force the material's magnetic spin to the desired direction using even larger currents to produce an opposite magnetic field.

      I suppose you could make all this parallel and get though put that way, but all this pulse generation is going to take lots of power if you are doing hundreds of thousands bits wide to make up for the latency..

      No, I don't see this stuff being used as cache. It will be too slow. Maybe for things that don't change often or require to maintain their content when powered off, but not cache.

      --
      "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
    7. Re:Will storage and memory eventually become one? by mentil · · Score: 2

      Indeed. If it were used for storage, it'd certainly cause a Space Oddity.

      --
      Corruption is convincing someone that the selfless ideal is the same as their selfish ideal.
    8. Re:Will storage and memory eventually become one? by Anonymous Coward · · Score: 0

      Microsoft would have to make some major changes to Windows if that happened. I'm always having to reboot to get some strange Windows behavior or another to stop. If rebooting didn't flush out the memory then the OS would be acting just as strangely after the reboot.

  3. Everything old is new again by Locke2005 · · Score: 2

    Anyone remember magnetic core memory? You used to be able to turn off the machine, then turn it on a week later have have it still be in the same state. We've just greatly improved the density of magnetic cores. (Sharp also had patents on magnetic memory chips.)

    --
    I've abandoned my search for truth; now I'm just looking for some useful delusions.
    1. Re:Everything old is new again by Anonymous Coward · · Score: 0

      Spintronics has been a dream for many years but the technology and the need has not been there to warrant the investment. Now that the world of computing meets its worsening energy crises the interest should be there again. Imagine instant on and off machines with non-volatile everything. Scalable, inexpensive MRAM would be a start, and a great one at it.

    2. Re:Everything old is new again by bobbied · · Score: 1

      Anyone remember magnetic core memory?

      Oh yes... I've used the stuff early in my career. It was part of the avionics package on a military fighter aircraft I was an engineer on. The amazing part was that stuff was nearly bullet proof, almost literally. I remember it was part of the flight data computer and provided a poor man's flight data recorder of sorts. A couple of times we had to read the contents from crashed aircraft to try and figure out what happened when the pilot was unfortunately unavailable to tell us what happened.

      I still remember having to visually inspect the stuff under a microscope to make sure the cores where not broken before getting out the test equipment to read the card's contents. Talk about mind numbing work.. But we could get a few seconds of flight data and that often was invaluable to determine why things went wrong.

      --
      "File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
  4. Requires further development by Drethon · · Score: 1

    "The new devices, with their low power consumption and high switching speed, could eventually be especially useful for devices such mobile computing, Beach says, but the work is still at an early stage and will require further development."

    So we don't yet know if it can replace existing hardware, but hopefully it will work out.

    1. Re:Requires further development by Anonymous Coward · · Score: 0

      True, as interesting as this may be, they should give up on investigating any practical uses. Maybe it will find it's way into some quirky product at Sharper Image.

  5. Want even lower power consumption? by Anonymous Coward · · Score: 0

    Ask Chuck Moore to design the chip in this fancy new process of yours. Go check out how much power his GreenArrays chips use, that you can buy today.

  6. more vaporware by Anonymous Coward · · Score: 0

    let me know when this shows up in a commercial product. more "research" that doesn't work outside the lab and goes nowhere.

    1. Re:more vaporware by ShanghaiBill · · Score: 1

      let me know when this shows up in a commercial product.

      If you are not interested in learning about leading edge research, then why are you reading a nerd website?

      more "research" that doesn't work outside the lab and goes nowhere.

      We have made huge strides in computing power and efficiency, year after year, because of the efforts of the very researchers you so flippantly denigrate.

      Go back to Facebook.

    2. Re:more vaporware by Anonymous Coward · · Score: 0

      all of those things you speak of are incremental improvements on existing technology. There hasn't been anything really groundbreaking since the transistor or jet engine. i'm still waiting on all those cold fusion and 3d holographic storage breakthroughs from the 1990s.

    3. Re:more vaporware by ShanghaiBill · · Score: 1

      all of those things you speak of are incremental improvements on existing technology.

      No they aren't. SSDs are not an "incremental improvement" over HDDs. They are a fundamentally different technology. There have been similar fundamental shifts in networking, and circuit board fabrication. High-K semiconductors, fractional-step lithography, and Fin-FET are all still based on transistors, but they have added up to revolutionary improvements in performance.

      i'm still waiting on all those cold fusion and 3d holographic storage breakthroughs from the 1990s.

      Scientific research doesn't work that way. Of course there will false hopes and dead ends. But only an idiot would obsess on those failures, while ignoring the cell phone in his pocket that is only made possible by profound breakthroughs.

    4. Re:more vaporware by Anonymous Coward · · Score: 0

      again everything you speak of is still based on the silicon transistor from 1947. solid state drives have been around for decades, cray offered one with the cray-2 in the 1980s. it's still transistor based. the transistors are just smaller and more numerous.

  7. Yay mining by Anonymous Coward · · Score: 0

    including a layer of cobalt where the magnetic changes take place, sandwiched between layers of a metal such as palladium or platinum, and with an overlay of gadolinium oxide

    cobalt... palladium or platinum... gadolinium...

    Holy crap. Where do you even get gadolinium?? Moria?

    1. Re:Yay mining by ShanghaiBill · · Score: 1

      Where do you even get gadolinium?

      Mostly from China, but gadolinium is also mined in the US, Brazil, Sri Lanka, India, and Australia. It is co-produced along with other rare earths.

      It is not particularly expensive, about 50 cents per gram, and a typical cell phone or computer based on this new tech would only use a few milligrams, costing less than a penny.

  8. Good point by SuperKendall · · Score: 1

    that's 3 orders of magnitude slower for the SSD.

    That's a great point, didn't consider latency. Even there though I wonder if some new storage technique like the one from the article may be able to close that gap, since it is an entirely different technology... but as you say that's a huge gap. I was more heartened by the somewhat less than an order of magnitude in terms of transfer speed.

    --
    "There is more worth loving than we have strength to love." - Brian Jay Stanley
  9. Can retain their magnetic properties by Anonymous Coward · · Score: 1

    For how long?

    And how far away do fridge magnets need to stay?

    1. Re:Can retain their magnetic properties by ShanghaiBill · · Score: 1

      For how long?

      They tested for 2000 cycles, and saw no degradation at all. So it is almost certainly reliable enough for SSDs. It is not clear if it is reliable enough for RAM, which would have to be good for billions of cycles.

      And how far away do fridge magnets need to stay?

      The effect is based on the movement of hydrogen ions. A refrigerator magnet would not be strong enough to interfere. But you likely wouldn't want to leave a strong magnet sitting on top of your phone for long periods of time. Even more so if your phone is hot, perhaps sitting in a car parked in the sun while running Bitcoin mining malware.

    2. Re:Can retain their magnetic properties by Ungrounded+Lightning · · Score: 2

      For how long?

      Geologic time (absent extreme temperatures or magnetic fields) wouldn't surprise me.

      Seabed minerals have retained the magnetization that they fossilized when they cooled below their curie point, creating a geologic record of Earth's magnetic field reversals in the spreading seabed. (That's how we know that/when the Earth's field reversed from time to time.)

      While this isn't the same structure, electron spins flipping in an environment conducive to them being stable is not something that tends to happen spontaneously.

      --
      Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
  10. Hohoho MIT never stops making is laugh by fubarrr · · Score: 1

    I congratulate the MIT for reinventing the bicycle... well, magnetic tunnel junction. Beg for grants and investor money more

    1. Re:Hohoho MIT never stops making is laugh by Drethon · · Score: 1

      I congratulate the MIT for reinventing the bicycle... well, magnetic tunnel junction. Beg for grants and investor money more

      There haven't been many inventions for decades (centuries?) that didn't reinvent some previous invention in some way that worked a little bit better. If that wasn't the case, there would be maybe one or two PHD graduates in each discipline, in the entire world every year.

    2. Re: Hohoho MIT never stops making is laugh by Anonymous Coward · · Score: 0

      By your logic, fire and the wheel are just incremental improvements on waiting for lightning to strike and hoping a big round stone or tree trunk is nearby when you want to push something.

      So you may not understand why we discount your opinion. But we do, and do.

  11. APK LIES AGAIN by Anonymous Coward · · Score: 0

    You were challenged to prove you've been "impersonated" beyond a single c6gunner post. You ran like the chickenshit you are. It's more likely that you made both posts because you're desperately craving attention for your garbage posts.

    1. Re:APK LIES AGAIN by Anonymous Coward · · Score: 0

      ZIP why did you try to downmod hide evidence of all the times APK fried you for your lies and fuckups https://news.slashdot.org/comm... ? Ashamed of your fails?

  12. IMPERSONATING me AGAIN? apk by Anonymous Coward · · Score: 0

    See subject: Hope you're RIGHT (considering I'm only sure hosts stop portsmash vs. Spectre/Meltdown) https://tech.slashdot.org/comm...

    You got caught impersonating me c6gunner (your name's the submitter signing "APK") https://linux.slashdot.org/com... as you ALTERED /.ers PRAISE of my work (not yours you don't even HAVE, lol).

    (You shouldn't throw stones when you live in a glass house boys - especially vs. me: RIGHT, ZIP? https://developers.slashdot.or... CAUGHT LYING TOO (you DO have a registered /. acc't. but STALK me anonymously instead - punk) https://news.slashdot.org/comm... )

    APK

    P.S.=> GROW UP weezils - you do it to yourselves trying to "take me on" & FAILING like you always do (especially on tech) + so then you start STALKING me by UNIDENTIFIABLE anonymous posts OR by IMPERSONATING me (weak BITCH tactics only a HOMO would do, lol)... apk

  13. This will never work by Anonymous Coward · · Score: 0

    This will never work, it has nothing to do with quantum physics.

  14. Thought Experiment: I5 Warp by Anonymous Coward · · Score: 0

    If you've missed any, please step back and read the earlier ones first.

    Switch on a torch.

    Postulate A: Mass isn't real
    Postulate B: the energy in light is also 'kinetic'
    Postulate C: Light bind force must be cyclical
    Postulate D: only 2 fundamental particles are possible
    Postulate E: the only force is electric
    Postulate E2: The binding force (Postulate C) is electric
    POSTULATE F: The speed of light is obvious
    POSTULATE G: Time is measured in spins
    POSTULATE H: All dipoles are equal, matter,even red and blue light
    Postulate I1: Donut Particles
    Postulate I2: Donut Particles are themselves dipoles
    Postulate I3: Anti-particles
    Postulate I4: Bigger particles twist and break
    Postulate J0: How light binds to matter
    Postulate J1: A Slit is a phase sorter
    Postulate J2: Gravitational lensing is just diffraction
    Postulate J3: Electron is a Donut Sandwich
    Postulate J4: Binding force is harmonic electric
    Postulate J5: Photons from Electrons
    Postulate K: How fast do forces propagate?
    Postulate L: Warp, Time Machine, Light squared computers

    ---------------------

    Postulate L: Warp

    Obviously the biggest consequence comes from Postulate G.
    If time is a function of these frequency F dipole spins. You could change this frequency and change the rate at which time passes.
    A slower F means a slower passage of time (and slower speed of light), a larger F means a faster passage of time.

    Testing a lengthy experiment? Put it in a time machine, speed up time, see the result quickly.
    Terminally ill? Enter a time machine, slow down time, wait for a cure.
    Want a super fast computer? Start the software, stick your laptop in a time machine, speed up time...result... superfast computer.

    Can it be done?

    Well let's focus on travelling as fast as light.

    ----

    Can we make matter travel at the speed of light?

    Pick up a torch, turn it on, you see that light? It was in the battery, as matter a few seconds ago. Now its light travelling at C. Quite literally it was a (2F donut - 2F anti-donut) dipole wrapper around a -ve monopole, and now it light. [Postulate J5]. Can we stop it again? Shine the light on a black body.... now its matter. Duh!

    So there, we've done it, matter travelling at the speed of light!
    Mission successful!

    ----

    OK, realistically, consider the two real problems here:

    1) To speed up F to F*2, you need to build some sort of oscillator that can induce waves at F*2.
    You cannot do this with electronics, electric force propagates infinitly fast (Postulate E), but the effect through dipole stuff propagates at the speed of light (Postulate k). Speed of light is W per spin. The electrons in your circuit would need to move the wavelength W*2 per spin, when it it only moving W per spin.
    Impossible. You couldn't even induce a spin of F*1.0000000000001, a tiny increase, because of the speed of light of the electrons in your circuit.

    2) The matter around your warp drive are oscillating at F. The whole universe around your warp drive is oscillating at F! And because its an electric force it is propagating instantly. The universe is far away, but the walls of your oscillator are far closer. Even if you moved the experiment far away from earth out into space, how can you find all that energy?

    To raise the spin to spin*1.5 and hold it, you'd need to counter all the binding energy in every dipole in your oscillator circuit, to sustain that 180 degree out of phase point. How much energy is in a few tons of equipment? Well if a few atomic level bonds contain a nukes worth of energy, we're two levels below that, a few billion billion billion times more than the energy in a nuke. That's a pretty amazing oscillator you've need there, more energy than this solar system.

    -----

    Lets lower our expectations here, and aim for some tiny increase F*1.000001 because that proves the principle and the energy might be within our reach and removes problem 2.

    Problem 1

    [BI

  15. Alternative strategies by Anonymous Coward · · Score: 0

    Strategy virtual:
    Can you make a shaped oscillating electric field? A net result of many electric fields that has a peak as far away as possible, and as high as possible.
    i.e. can you solve the problem 2 by making the oscillator a virtual oscillator far far away, with only the field propagating to the target?

    Strategy clear the wake
    Look at the addendum comment to Postulate K. Gravity propagates through dipolar phase changes, which means it travels at C through matter. The addendum points out that if there is no dipole in a wavelength W, then there is no corresponding phase change. In other words the electric field skips one cycle.
    I suspect your gravity wave experiment has some sort of physics fixup to account for/obsfucate this effect, which is why you think it travels at C through thin space, when it travels faster than C.

    But here is another strategy, clear one wavelength W of space between you and the gravity source. Now you are 1 phase change nearer your destination.

  16. Why was this a breakthrough again? by OneOfMany07 · · Score: 1

    Had to look in TFA to figure out what was special about this. Looks like this paragraph...

    "But spintronic technology suffers from its own limitations. One of the biggest missing ingredients has been a way to easily and rapidly control the magnetic properties of a material electrically, by applying a voltage. Many research groups around the world have been pursuing that challenge."

    Wish that'd been in the summary of the article. I'd heard of spintronics years ago, so that isn't such a useful buzz word...