Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×

Intel Previews Potential Replacement for Flash Memory 131

Posted by Zonk from the welcome-to-the-world-of-tomorrow dept.
GeeksAreSexy writes "Eweek has an article up about the invention of a new kind of nonvolatile memory technology that could one day replace traditional flash memory. Unlike traditional flash memory, chips using this new technology will be able to execute code with performance, and sustain millions of read/write cycles without dying." From the article: "This is a case in which 'Necessity is the mother of invention' is very true. We were forced to look for something else, completely different. That's why we decided to invest in PCM ... There are definitely limits to what you can do with our current flash methodology. There needs to be a complete quantum leap somewhere along the line to push everything forward. We believe PCM are going to be that quantum leap."
This discussion has been archived. No new comments can be posted.

Intel Previews Potential Replacement for Flash Memory More

Intel Previews Potential Replacement for Flash Memory

Comments Filter:

  • Damn... (Score:4, Funny)

    by tonigonenstein ( 912347 ) on Friday September 29, 2006 @09:21AM (#16243865)
    ... and here I tought naively we could kiss goodbye to Macrobe Flash.
  • Just my 2c worth but I remember seeing this in a story (from Samsung) using the same technology at least 12 months ago Is this a reissue?
    • No...it's just that Intel's found out about it now :)

    • Re: (Score:2, Funny)

      by Gospodin ( 547743 )

      Impossible for three reasons:

      1. Slashdot only prints the most current, up-to-the-minute stories you won't find anywhere else.
      2. Slashdot never prints duplicates.
      3. Intel never "borrows" technology from competitors.

      You're welcome.

    • Re:And this is NEW? (Score:4, Informative)

      by cyfer2000 ( 548592 ) on Friday September 29, 2006 @10:52AM (#16245067) Journal

      While the whole thing is a little more complicated. Ovshinsky [wikipedia.org] was the first one to get patent on this area, and he opened a company named Ovonics [ovonic.com]. Then Ovonics created a company named Ovonyx [ovonyx.com] with a cofounder of Micron. Ovonyx is focused on the Phase Change RAM [wikipedia.org] while Ovonics keeps working on things like Fuel cell, Solar cell, batteries...

      Gordon Moore of Intel was also one of the early researchers on the area of Phasse Change RAM. In 2000, Intel invested some big money into Ovonyx and get the license of Phase Change RAM from Ovonyx. Samsung licensed the Phase Change RAM from Ovonyx later.

  • Another one already? (Score:4, Interesting)

    by Ant P. ( 974313 ) on Friday September 29, 2006 @09:22AM (#16243899)
    So is this better or worse than that other "flash replacement" memory we heard about on /. the other week? You know, the one that's supposedly got the best parts of DRAM, hard disks and flash all in one?
    • Mixing hard disk, flash, and dram in one device sounds more like a hard disk replacement, since it still has moving parts and would really just improve access times to data stored on the platters.

      • Re: (Score:3, Funny)

        by mgblst ( 80109 )
        Yeah, I guess when he said parts, he didn't mean physical parts. What do you think?

  • Nirvana (Score:3, Funny)

    by Red Flayer ( 890720 ) on Friday September 29, 2006 @09:24AM (#16243907) Journal
    "PCM is like a super set of NOR or NAND flash," Doller said. "It's almost nirvana for an engineer. It reads fast, writes fast--it does everything faster."
    Nah, if it were nirvana for an engineer, it would do everything just as fast as it needs to, and no faster.
    • If it were nirvana, it would include a mini hologram of scotty inside working hard to deliver your data when you need it no matter what the marketing information says.

  • Quantum Leap (Score:5, Funny)

    by tygerstripes ( 832644 ) on Friday September 29, 2006 @09:24AM (#16243917)
    There needs to be a complete quantum leap somewhere along the line to push everything forward. We believe PCM are going to be that quantum leap.
    You mean... like... a leap so small that it's impossible to make a conventional leap any smaller, and measuring and predicting effects on such a tiny scale are so experimental and imperceptible that they require a unique perspective of the laws of nature in order to make any sense of them?

    Hardly news then, right?

    • Re:Quantum Leap (Score:5, Funny)

      by jeffmeden ( 135043 ) on Friday September 29, 2006 @09:43AM (#16244113) Homepage Journal
      Trapped in the past, Intel finds themselves leaping from technology to technology, putting things right, that once went wrong and hoping each time, that their next leap will be the leap home.

      that kind of leap...

    • Re: (Score:1, Offtopic)

      by steveo777 ( 183629 )
      I think you're thinking of epsilon... which I have no clue how to type into a /. comment.

      In my Calc II class in college we had a couple TA's. One was there to help students and the other was there to grade papers. They butted heads a lot becuase the TA that helped the students did it in a way that pissed off the grader (didn't show enough work or something like that on the homework). Well, I'd frequently get knocked off a few points because I didn't show a step here and there (usually nitpicky stuff like

    • Re: (Score:2)

      by pacc ( 163090 )
      We believe PCM are going to be that quantum leap

      Almost correct, a Quantum Leap may take you anywhere,
      like through walls, to unprecedented performances.

      Though in the future when we know the results of this
      invention and the probability distribution breaks down
      it may well show that the quantum leap was a big leap
      backwards or nowhere at all.

    • Re: (Score:2)

      by dylan_- ( 1661 )
      Well, I know you were just being funny but, lest anyone be mistaken, the physical size of a quantum leap is not the aspect that is being utilized in the metaphor. It is the direct change from one state to another without any intervening steps. You might consider it as opposed to an "evolutionary" technology which would be an improvement on previous work.

      However, anyone using the phrase "evolutionary quantum leap" [google.com] should, of course, be shot.

  • Similar to CD-RW? (Score:2, Insightful)

    by cperciva ( 102828 )
    Quoth the article:
    PCM chips use the same material, chalcogenide, that's used inside to store data in a rewritable optical discs.

    For a system designed to "sustain millions of read/write cycles", this seems a bit strange -- last I heard, CD-RW disks were limited to a few hundred rewrites, never mind millions.

    • Re:Similar to CD-RW? (Score:5, Insightful)

      by mgblst ( 80109 ) on Friday September 29, 2006 @09:44AM (#16244123) Homepage
      You realise that CPUs user the same material as most beaches, but they still manage to give them many more FLOPS than your average beach. It is almost as if it is more important the way the material is used - nah, that can't be true. If one product is made out the same material as another product, then it must be exactly the same!!!
      • Try reading TFA:

        PCM chips use the same material, chalcogenide, that's used inside to store data in a rewritable optical discs. But instead of using a laser to change the properties of the material and thus create the zeros and ones that make up data, the chips use electricity that flows through a resistor. The resistor heats up and does the job of the laser, changing the material's properties to represent a zero or a one.

        This sounds exactly like the phase change in CD-RW, albeit done in smaller scale

        • Re:Similar to CD-RW? (Score:5, Insightful)

          by saider ( 177166 ) on Friday September 29, 2006 @10:17AM (#16244533)
          but there's no fundamental technical difference as far as I can see

          Except that one is changed with a laser and the other is done electrically?

          The laser is probably more powerful than it needs to be because it needs to pass through a (relatively) dirty lens, several mm of air, and a layer of plastic before altering the material. in order to do this reliably, they overpower the laser so that it can achieve the effect. The tradeoff is that the excess power wears the material out faster.

          Now the material is integrated into a chip and uses simple thermal conduction instead of radiation to achieve the effect. The distances are much smaller and the environment is much more controlled, which means that you do not need to overpower the devices. This results in reduced wear, which means a longer life.

          As the GPP said...
          "It is almost as if it is more important the way the material is used".

          • Now the material is integrated into a chip and uses simple thermal conduction instead of radiation to achieve the effect. The distances are much smaller and the environment is much more controlled, which means that you do not need to overpower the devices. This results in reduced wear, which means a longer life.

            This is what I meant when I said the process is tweaked to be much better. But the difference is just a matter of degree. It's hard to imagine that you can take something that wears out after 100

      • Re:Similar to CD-RW? (Score:4, Funny)

        by CortoMaltese ( 828267 ) on Friday September 29, 2006 @10:11AM (#16244439)
        You realise that CPUs user the same material as most beaches, but they still manage to give them many more FLOPS than your average beach.
        You're comparing apples and oranges. When comparing CPUs and beaches, instead of flops [wikipedia.org] you really should compare flip-flops [wikipedia.org] and flip-flops [wikipedia.org].

      • Re:Similar to CD-RW? (Score:4, Informative)

        by dhovis ( 303725 ) * on Friday September 29, 2006 @10:15AM (#16244487)
        You realise that CPUs user the same material as most beaches

        Ummmm.....No. Beach sand is mostly silicon dioxide, whereas computer chips are fabricated starting from wafers of very pure silicon.

        Diamond (pure carbon) and carbon dioxide don't have similar properites either.

        Sorry to get pedantic, but I'm a materials scientist, and it really pisses me off when people get these things mixed up. It is even worse when people confuse silicon (the base material for computer chips) with silicone (a polymer material used in caulking and breast implants).

        • It is even worse when people confuse silicon (the base material for computer chips) with silicone (a polymer material used in breast implants).
          Does that explain why this idiot [kevinwarwick.org] is always having computers implanted into his flesh?

        • The useful part of a CPU is mostly SiO2 with some metal and polycrystaline silicon and a small amount of doped single crystal silicon.

          Now adays many of the SiO2 in those fancy CPUs has been replaced by SiLK and maybe MMSQ. SiLK is cross linked organic stuff, the real structure has not been revealed yet. MMSQ, IMHO, is a kind of silicone.

          • Re: (Score:3, Informative)

            by dhovis ( 303725 ) *

            OK, I'll give you a big chunk of quartz (SiO2), and you make a transitor out of it.

            There are a lot of materials that go into a computer chip, and SiO2 plays an important role (mostly as a gate dielectric), though as you say, other materials have started to be substituted in. But to say that SiO2 is the useful part? You can't make a transistor out of SiO2 only, it doesn't conduct electricity.

            Take a look at the MOSFET illustrated here: http://en.wikipedia.org/wiki/Image:Lateral_mosfet. svg [wikipedia.org]. The oxide

            • Re: (Score:1)

              by anishm ( 770357 )

              SiO2 is only used because it is easy to make by oxidizing silicon and it does not react with silicon. For a long time, it was "good enough". Low-k materials would be better, and that is why you hear about them more.

              Electrically at least high-k gate diectric is supposed to be better. I think this is largely because it allows better gate control of the channel without having to make the gate diectric so thin that electrons "tunnel" through it - i.e make quantum leaps across the gate barrier :). E.g. here

        • Ummmm.....No. Beach sand is mostly silicon dioxide, whereas computer chips are fabricated starting from wafers of very pure silicon.

          And yet that pure silicon came from... I dunno... silicon dioxide, maybe?

          If you're a materials scientist, you should know better than to think polySi exists in any usable quantities in nature, and you should know that the doped polySi used in wafer production comes originally from reduction of Si02 by burning with a carbon fuel (like wood) at very high temperatures.

          • Re: (Score:2)

            by dhovis ( 303725 ) *

            And yet that pure silicon came from... I dunno... silicon dioxide, maybe?

            Steel is made by reduction of iron ore, but that doesn't make steel and iron ore the same material.

            And anyway, the polysilicon made from the reduction of SiO2 (quartz) still needs to be purified by zone refining before it can be used in the manufacture of computer chips. Then single crystals of silicon are usually grown using the Czochralski method and sliced into wafers. So beach sand is quite few processing steps away from the

            • Regardless, silicon chips are made from the same material that sand is primarily composed of. The manufacturing process starts with raw materials.
        • Do you have a model of Perovskite on your desk too? :)
      • The CPU is made out of Silicon with small controlled amounts of impurities. The beach is made out of Silicon Dioxide with small uncontrolled amounts of impurites.

        Silicon is to SIicon Dioxide as Hydrogen is to Water.

        Frankly, I have a few doubts myself about anything that uses any part of the technology associated with CDR or CDRW. One hopes the technology works a LOT better when it is integrated onto a chip.

        I'd also point out that maybe 15% of these magic new technology announcements ever make it into

    • Re: (Score:2)

      by Ant P. ( 974313 )
      That's because microchips don't double as centrifuges.
  • The main problem that I forsee is that it seems to me like os many different companies are pursuing their own form of what will be the successor to flash. What kind of implementation will this have, and is it going to negatively affect the customer by increasing the number of devices that are incompatible?

    • Re: (Score:1)

      by Jessta ( 666101 )
      I'm sure the customers don't evening know that their data is stored on magnetic disks. It doesn't matter how the data is stored,it only matters how the device communicate with other devices. Device with this technology will probably just use the standard USB mass storage device driver.

    • The main problem that I forsee is that it seems to me like os many different companies are pursuing their own form of what will be the successor to flash. What kind of implementation will this have, and is it going to negatively affect the customer by increasing the number of devices that are incompatible?

      It will have zero impact on computer users.

      Engineers designing with these parts will have to implement some kind of controller that sits on either the processor's local bus or a mezzanine bus, and even

    • What kind of implementation will this have, and is it going to negatively affect the customer by increasing the number of devices that are incompatible?
      I suspect that compatibility is not going to be that big of an issue, as long as they support similar interfaces, just like flash vs. disk-based USB drives now.
  • PCM? What happened to the nanotubes? I wan't my petaflop-performance-lighter-than-air-stronger-tha n-steel-elevator-into-space-that-will-store-my-pr0 n-collection to replace my flash thumb drive.

    (sigh) So many empty promises.
  • that they were manufacturing this new memory out of the recycled parts from millions of discarded RDRAM chips.
  • ... I had a faint hope of Intel having a Flash Player 9 for Linux and AMD64 ready. Yes, I feel stupid now, but I'll call it preoccupation.
    • That's what I thought when I saw this headline on Geeks Are Sexy [blogspot.com] yesterday.
    • You shouldn't feel stupid. Blame the ed like everyone else on /. does.

      For one thing it actually isn't your mistake. There a difference between Flash and flash. It's an error in the headline.

  • would people please stop using "quantum leap" a say "giant leap"??
    pretty please?

    • Re: (Score:2)

      by Aladrin ( 926209 )
      Maybe 'truth in advertising' has taken an unexpected entry into real advertising. This probably really is literally a 'quantum leap' for memory.

    • Re: (Score:3, Informative)

      by Tim C ( 15259 )
      I'm afraid >a href="http://dictionary.cambridge.org/define.asp? k ey=64758&dict=CALD">it's too late.

      Besides which, the "original" quantum leap isn't "the smallest possible leap", it's a movement of a particle that should be impossible - iirc, it's either when a particle moves from point A to point B without passing through the intervening space, or when it crosses a potential barrier that (classically speaking) it's impossible for it to cross.

      The idea being not that it's a giant leap forward, but t

      • Re: (Score:1)

        by nasch ( 598556 )

        Besides which, the "original" quantum leap isn't "the smallest possible leap", it's a movement of a particle that should be impossible

        Well I'm no physicist, but that doesn't sound right to me. Isn't a change from one quantum to another specifically a step in energy level? For example, a laser excites particles (electrons?) from one energy level to a higher one, then they emit photons when they return to the lower energy state. This is a quantum change because it's a discontinuous step directly from one e

    • Re: (Score:3, Informative)

      by realnowhereman ( 263389 )
      Will people stop confusing the word "quantum" with "small"?

      A quantum leap does not mean "the smallest possible leap"; it just so happens it is the smallest possible leap. However, it means a jump from one discrete energy level to another, without passing through anything in between.

      To me, that makes "a quantum leap in technology" a perfectly resonable metaphor. Albeit completely overused by marketing droids.

    • ...was certainly a "quantum leap", not a small step. I don't know the etymology of the phrase, but I always thought it had something to do with the revolutionary transition from classical to modern physics, and all the resulting technology that stemmed from that. I could be wrong of course, or right. I won't know until I measure.

    • Quantum does not mean "a small amount" it means "amount" as in a discrete amount rather than a continuous change.

      Quantum is used in quantum mechanics because classical chemistry said that electron energies (for example) scale continuously, but experimental work shows that there are discrete energy jumps, i.e. there are fractional energy levels that are not phycically possible for a given system. The language itself doesn't care that this happens on the Angstrom level or the kilometer level. A "quantum leap"

    • Re: (Score:2)

      by mnmn ( 145599 )
      Well if you know where you leap from, you can never know how big the leap is.

  • Headline Changed (Score:3, Informative)

    by abscissa ( 136568 ) on Friday September 29, 2006 @10:01AM (#16244303)
    The headline has been changed clandestinely and the word "memory" has been added... in case you are wondering about some of the comments before this one.

    Like many people here, I too saw the headline and thought some replacement for macromedia flash was on its way...
    • i saw the headline and was wondering why a pulse code modulation interface would have anything to do with the underlying storage technology, and why it would be a good idea in the first place.
    • clandestinely [webster.com]: marked by, held in, or conducted with secrecy.
      • What's your point? The headline was changed without any mention of it... therefore it was conducted with secrecy?
        • You wrote, "The headline has been changed clandestinely ...", I didn't.

          The word you used implies that something sneaky was going on when in fact the headline change was a simple clarification.

          *That's* my point, see?
  • The chief technology officer for Intel's Flash Memory Group, Ed Doller, is the guy quoted as calling for a quantum leap. He's probably a marketing guy, not a technologist. No harm done because CTO tends to be more of a marketing position than a tech one. You would hope for a stronger mind in that position though.

    Or perhaps he's typical of the slashdot posters who don't understand basic quantum theory. Check wikipedia [wikipedia.org] under the vernacular usage heading.

    Personally, I like the phrase "quantum leap" be
  • Why are we stuck recycling an acronym already done to death by analog-to-digital and Mexican communists? [wikipedia.org]

    I say, bring back twistor memory and bubble memory. Sure, they worked like crap, but their names were just so much cooler!

    • I say, bring back [...] bubble memory

      [Checks watch] Yep, it's been twenty years, time to trot that tech out again. Maybe they've figured out how to double-clock it now, so we can get access times down to single-digit milliseconds!

      Actually, it's worth noting that NEC got access time for bubble memories down to 16.5ms, roughly a third the speed of disk drives at the time. The only machine I ever used that had bubble memory was a TI 765 portable data terminal. It had 20K, which allowed the user to enter te

  • No RAM, just a BFO RAMdisk and a big on CPU cache.

     

  • Millions of read/write cycles... (Score:4, Interesting)

    by joto ( 134244 ) on Friday September 29, 2006 @11:11AM (#16245371)

    [snip] execute code with performance, and sustain millions of read/write cycles without dying.

    Wow! That means that in the worst case, it will last SEVERAL seconds!!!

    (Wouldn't it be better to have something like trillions of read/write cycles, so we know it will at least last a few years?)

    • Re: (Score:2)

      by suggsjc ( 726146 )
      Why stop there. Why not just go ahead an make it an infinte number of read/write cycles? That way it would last SEVERAL years. But then again, how long is technology like this supposed to stick around? Once something hits mainstream usage, there is already something bigger/better/faster/whatever that will come along and make it obsolete. Its a viscious cycle.

      All of that to say, we take steps forward to keep from moving backwards.

    • Re: (Score:2)

      by repvik ( 96666 )
      Wow! That means that in the worst case, it will last SEVERAL seconds!!!


      If wear-leveling isn't used. Which is rare. Even so, it's a great improvement from the current status with ~100k cycles.

    • Gotta love it when trolls like the parent get modded up. Had the article mentioned trillions of write cycles, he'd be asking for quadrillions.

      According to wikipedia:
      "most commercially available flash products are guaranteed to withstand 1 million programming cycles" -- the days of 10,000 write-cycle devices are long over.
      http://en.wikipedia.org/wiki/Flash_memory [wikipedia.org]

      Additionally, "Wear Leveling" spreads writes out over the whole of the device, greatly increasing MTBF
      http://en.wikipedia.org/wiki/Wear_levelling [wikipedia.org]

      If

      • Re: (Score:2)

        by joto ( 134244 )

        Gotta love it when trolls like the parent get modded up. Had the article mentioned trillions of write cycles, he'd be asking for quadrillions.

        Of course I would. What I really want is unlimited write cycles. I dislike things that fail even when used according to manufacturer specifications. But with trillions of write cycles, you would have a few years of use, even in the worst case (assuming this new ram is about as fast as dram). With quadrillions of write cycles, it would be very unlikely that the ram w

        • What, so hard drives are infalliable now? How many write cycles can your average drive handle? If you "dislike things that fail even when used according to manufacturer specifications", why are you so stuck on hard drives to begin with? They're the *most* failure prone component in PCs today.

          Yes, flash does suck on the MTBF metric -- but so do hard drives! It doesn't matter wether flash can handle a billion write cycles or 17 quadrillion -- it only really matters if flash drives have a comperable or bet
  • An excellent regurgitation of this whole thing is over at mymac.com

    SOLID STATE MACINTOSHES?

    . . . PCM chips do seem to be the long term replacement for flash memory chips, which is why you are reading this blog at MyMac.com. Apple Computer has a vested interest in INTEL and its advances, and we all know, if only subconsciously, that all computers some day will be driveless solid state devices with no moving parts at all (right?). Therefore, it is just a matter of time, because with the new PCM volatile/soli

  • A Lot of Potential, But a Long Way to Go (Score:3, Interesting)

    by organgtool ( 966989 ) on Friday September 29, 2006 @11:22AM (#16245623)
    According to Wikipedia, PCM [wikipedia.org] has the potential to squeeze a terabyte into one square inch. However, even a few gigs is enough to make a huge difference. PCM could be used to store the operating system and application files as well as for swap. Since PCM performs similarly to DRAM, it would be like having all of your applications loaded into memory at all times. And since this memory is non-volatile, going into and out of hibernation mode would be almost instantaneous and it would not use any power while in hibernation mode. It would also having the following benefits:
    • Lower power consumption since this has no moving parts and your computer would only need to spin the hard drive if the user is requesting a document file
    • Less noise since the hard drive could remain powered down if the user was not currently accessing documents
    • Less heat which would reduce the number of case fans required
    • Since it uses less power, laptops would get more life out of their batteries

    I can see a day where this memory is used in place of DRAM and application files are permanently stored in memory even when the system is off.

    • Re: (Score:2)

      by mnmn ( 145599 )
      Forget 'potential', MRAM is already there and even in production (although its expensive). It doesnt offer the same code density but its here now, and its reliability is more like DRAM than like flash.
    • The operating systems would probably have to be modified to make sure that it spreads its writes around evenly. I know there is a concept of "wear leveling", but I think in order for that idea to full its promise, it shouldn't involve a lot of files that aren't rewriten often along side with a lot of files that are rewriten often. Otherwise, it's effectively just "wear leveling" the free space + often changed files. Swap files are of particular concern, those files currently don't seem to change block lo

  • watch your metaphors (Score:3, Insightful)

    by oohshiny ( 998054 ) on Friday September 29, 2006 @11:31AM (#16245821)
    I think when people talk about "quantum leaps", they are forgetting two things: (1) quantum leaps are usually tiny, and (2) it is unpredictable when they happen. Is that the metaphor Intel wants?
    • "Quantum leap" also means a change to a completely new level - i.e not an incremental or minor change. Which is what they mean here.
    • The term "Quantum leap" does not refer to the size of the leap. It refers to the ability of particles at that level to go from point A to point B without crossing the intervening space. The metaphor is of making a jump from one place to another, bypassing all the intervening steps.
  • It's necessary to invest in technologies such as PCM because flash memory will eventually hit a wall in which it can no longer scale with silicon manufacturing.
    Does this have something to do with the way transistors work?
  • PRAM has been in development for years, and intel is nowhere on the map until now... too many "new memory" types and who knows which will win out... FeRAM, CRAM, MRAM are just a few that immediately comes to mind...

  • WTF (Score:2)

    by MS-06FZ ( 832329 )
    What the fuck does "execute code with performance" even mean? Does it imply good performance, or just some arbitrary kind of performance? If I give it a chance, will it do a hip-hop dance?
  • Is this *really* a quantum leap [wsu.edu]?

Slashdot Top Deals

If all else fails, lower your standards.

Close