Forgot your password?
typodupeerror
Data Storage IBM Hardware

Texas A&M Research Brings Racetrack Memory a Bit Closer 55

Posted by timothy
from the a-few-bits-really dept.
MojoKid writes "IBM is one of a number of companies working on a next-generation storage memory project and a recent discovery at Texas A&M University is a step forward for the company's racetrack memory. Racetrack memory relies on a nanowire arranged perpendicular to the chip. Current pulses across the nanowires allow data to be shifted as necessary. In theory, racetrack memory could be the Holy Grail of storage, capable of replacing both traditional hard drives and SSDs simultaneously. Racetrack memory could solve multiple problems and commercial implementations could offer hard drive-level density. Performance and reliability would both be far superior to today's SSDs. To date, IBM has demo'd a three-bit racetrack configuration. It's a start, but it's far from a shippable product at this juncture." What the A&M researchers have come up with is "a way to pulse the current much more efficiently and quickly."
This discussion has been archived. No new comments can be posted.

Texas A&M Research Brings Racetrack Memory a Bit Closer

Comments Filter:
  • by SuricouRaven (1897204) on Wednesday December 01, 2010 @07:59AM (#34403014)
    Somehow I am reminded of the old mercury delay line memory.
    Also, first?
    • Re: (Score:3, Informative)

      by splutty (43475)

      You must've clicked through to the original article, where that's been modded up quite a bit :)

      But yes. It has similarities. The main difference as far as I can see is that this is much more '3D'. I sort of imagine a carpet waving around (funny mental image, do we now need tiny kittens to scratch their nails on that?)

      But the option of being able to store multiple bits no a relatively small footprint is of course the most interesting one, although I wonder about the heat dissipation or production of these so

      • Re:Delaylineish. (Score:4, Informative)

        by SuricouRaven (1897204) on Wednesday December 01, 2010 @08:06AM (#34403078)
        A more important difference is that the old delay lines had the bits constantly moving along - when you wanted to access one, you just waited for it to pass by the transducer. This model has the bits hold still until they are made to move back or forth along their track, like a train moving one-carriage-at-a-time past a loading crane.
  • by careysub (976506) on Wednesday December 01, 2010 @08:01AM (#34403032)

    Brings back the magnetic bubble memories [decadecounter.com]!

    • Themes for research in EECS tend to have about 25-35 year cycles while being updated with the technology du jour at each peak. I think it has something to do with the dying out/retiring of academic advisers and industrial lab directors who think a field is "mined out". Bubble memory is about that old. Mercury delay lines twice that. On the other hand, I get the distinct impression that most of the performance findings will show the same pros and cons of any cyclical memory system. It really points out

      • by careysub (976506)

        Themes for research in EECS tend to have about 25-35 year cycles while being updated with the technology du jour at each peak. I think it has something to do with the dying out/retiring of academic advisers and industrial lab directors who think a field is "mined out". Bubble memory is about that old. Mercury delay lines twice that...

        Magnetic bubble "racetrack" memory seems a much closer parallel than the delay lines. MBM had some things going for it - solid state, non-volatile and extremely rugged. So mostly it got used in niche military applications. Introduced in the 1970s it was "revived" in 1989 with the U. S. Army's AN/UGC-144 portable battlefield communication terminal. At the time it was said to be "the first time a high-volume, off-the-shelf magnetic bubble memory system will be a component in a production model military comput

    • Memory in UNIVAC [wikipedia.org]. Then Forrester perfected magnetic core memory and IBM magnetic disk memory.
  • by LostCluster (625375) on Wednesday December 01, 2010 @08:19AM (#34403146)
    Are the bits named Win, Place, and Show?
  • by Anonymous Coward

    The Kingston ladies sing this song.
    Do-dah, do-dah.
    The racetrack memory's five mils long.
    Oh, do-dah day.

    Goin' to write some ones.
    Goin' to see them stay.
    I bet my money on some Corsair DIMMs.
    Somebody boot up the Cray.

  • by digitaldc (879047) * on Wednesday December 01, 2010 @08:33AM (#34403250)
    Hopefully, the motto of this new storage device will not be "I lost EVERYTHING!"
  • by neo12 (1892318)
    Well...wait till guys from the quantum memory research teams publish their research.
  • I already have enough problems with users being unable to distinguish between memory (RAM) and storage (HDD/SSD). I will now have to deal with a type of storage called Racetrack Memory?!

    This is only a half-hearted joke. Does this new form of storage/memory obviate the need for RAM by simply allocating one of the racetracks to that type of storage/retrieval?

    • let's just bring back old-fart terminology. Primary storage and secondary storage. Your HDD and SDD and USB Memory stick and Racetrack are all secondary storage. RAM is primary storage.
      • by gman003 (1693318) on Wednesday December 01, 2010 @09:56AM (#34404074)
        Actually, the predicted speeds for racetrack memory are rather close to RAM speeds. It could be possible to operate it as primary storage - it would be a bit sluggish for calculations, but it would eliminate the need to read in chunks of files to be operated on. Given the increasing size of L3 cache, it could be possible for racetrack memory to replace SSDs and RAM while pushing hard drives into the "long-term storage" role, and have the L3 cache take the role of RAM.

        This isn't a firm prediction - I'm not even sure racetrack memory will come to anything, but if it does, "the death of RAM" is entirely possible.
        • Actually, the predicted speeds for racetrack memory are rather close to RAM speeds. It could be possible to operate it as primary storage - it would be a bit sluggish for calculations, but it would eliminate the need to read in chunks of files to be operated on. Given the increasing size of L3 cache, it could be possible for racetrack memory to replace SSDs and RAM while pushing hard drives into the "long-term storage" role, and have the L3 cache take the role of RAM. This isn't a firm prediction - I'm not even sure racetrack memory will come to anything, but if it does, "the death of RAM" is entirely possible.

          I'll agree the death of RAM is certainly possible if its fast enough. However, it will likely have a lot better life with less moving parts than a traditional hard drive. So I would see traditional hard drives definitely going away - SSDs will do that by themselves once they get cheap enough even without Racetrack memory taking on hard drives and SSDs. So I could see the long-term storage going to SSDs or Racetrack memory-based hard drives; but traditional hard drives will certainly die at some point. It's

  • Will on-line betting be legal?

Recent investments will yield a slight profit.

Working...