Slashdot Mirror

← Back to Stories (view on slashdot.org)

Micron Kicks Off Mass Production of 12Gb DRAM Chips (anandtech.com)

Posted by BeauHD on from the new-and-shiny dept.

Micron is now producing its first LPDDR4X memory devices using its second-generation 10nm-class process technology. "The new memory devices offer standard LPDDR4X data transfer rates of up to 4.266 Gbps per pin and consumes less power than earlier LPDDR4 chips," reports AnandTech. From the report: Micron's LPDDR4X devices are made using the company's 1Y-nm fabrication tech and feature a 12 Gb capacity. The manufacturer says that its LPDDR4X memory chips consume 10% less power when compared to its LPDDR4-4266 products; this is because they feature a lower output driver voltage (I/O VDDQ), which the LPDDR4X standard reduces by 45%, from 1.1 V to 0.6 V. Micron's 12 Gb (1.5 GB) LPDDR4X devices feature a slightly lower capacity than competing 16 Gb (2 GB) LPDDR4X offerings, but they are also cheaper to manufacture. As a result, Micron can offer lower-cost quad-die 64-bit LPDDR4X-4266 packages with a 48 Gb (6 GB) capacity and a 34.1 GB/s bandwidth than some of its competitors.

52 comments

  1. Re: Eliminate the WEAK. by Anonymous Coward · · Score: 0

    They're not centred. Weak. Time to eliminate yourself.

  2. Rowhammer? by Anonymous Coward · · Score: 0

    Are they fixed yet?

    1. Re:Rowhammer? by GabeGhearing · · Score: 2

      Yes and no, the LPDDR4 JEDEC(rather than vanilla DDR4) has TTR to mitigate Rowhammer... but support in memory modules is optional.

    2. Re:Rowhammer? by GabeGhearing · · Score: 1

      Oops, meant TRR(targeted row refresh), not TTR. https://en.m.wikipedia.org/wik...

  3. By an amazing coincidence... by Anonymous Coward · · Score: 1

    So did Jinhua Integrated Circuit Company.

    1. Re:By an amazing coincidence... by Anonymous Coward · · Score: 0

      Obligatory US reaction.

    2. Re:By an amazing coincidence... by Anonymous Coward · · Score: 0

      No way would I COPY 12 when 16 is here and now. With new greenfield/machines, go for 16's, and when they get worn out, drop them to 12's.
      Suffice to say Jinhua probably said rubbish - we will go the Samsung way, and even pay Samsung something. They now KNOW Micron has no easy fixes, and know they can make fab gear it will be the same or worse. so work with TMC, and plan to undercut Micron on price knowing they have little wriggle room.
      Unless Micron is putting out clever disinformation, and has 20's ready to go.

  4. Re: Eliminate the WEAK. by Anonymous Coward · · Score: 0

    You lost the ascii art battle

  5. What is "1Y-nm" ... here's an article by NothingWasAvailable · · Score: 5, Interesting

    As someone who worked in semi-conductor CAD, 1Y-nm confused me.

    I found an article in EE Times that explains is using 19 nm to 10 nm as three nodes at 1X, 1Y, 1Z, with X, Y, and Z to be defined later.

    1. Re:What is "1Y-nm" ... here's an article by NothingWasAvailable · · Score: 4, Informative

      https://www.eetimes.com/docume...

    2. Re:What is "1Y-nm" ... here's an article by Anonymous Coward · · Score: 1

      they mean the x and y have scaled down so the new process is 0.4 x 0.5 y, but only 0.9 z , so it didnt scale linearly like older processes 1-1

      7 nm as defined by marketting is only describing the smallest feature, rather then the average size of the node ... i dont believe the processes ever scaled linearly, but is now more disjoint ..... as the taper from the print feature will define the spacing between the elements

  6. Told you it would be useful by Anonymous Coward · · Score: 0

    If the time taken to propagate electric force is a function of the time taken for 1 dipole oscillation..... i.e. a phase change
    And if you can bind dipoles at 1:1 1:2 1:3 1:4 to other dipoles

    And if matter binds at 1:1 to other matter

    Then you could 'theortically' build a circuit where all matter dipoles spin twice as face, electrons, protons, molecules... all made of dipoles spinning twice as fast like a form of stationary light. Except its matter.

    Cut the phase changes to half, and electric force will propagate through such a circuit at 2C!

    Sure I make it sound easy, by avoiding little things like making it, and isolating it so that 1:1 maximum doesn't mess it up..... but ....

  7. What is the magic number from? by Anonymous Coward · · Score: 0

    I cheated, I swapped one magic number for another magic number. Replacing C with W derived from F.

    The basis that all dipoles even out to be coherent, due to their interactions, like this:
    https://www.youtube.com/watch?v=5v5eBf2KwF8

    But that gives us a magic number 'F". Where did F come from?

    Well from the peasoup model (my simulation of dipole stuff), F simply depends on the start conditions. Spread out some random monopoles, set the model off and you get a different F each time!

    i.e. F is arbitrary for any universe and so the speed of light is also arbitrary!

    THIS DOES NOT MATTER!

    Imagine two universes that began as monopole stuff, big bang, dipole stuff.... their F would be different.
    If those two universes (F1, F2) merged the value of F would change to some balance with the new merged F1 and F2.

    Postulate G explains the nature of time, which helps understand why the value of F is irrelevant (and best defined as the value 1)!

    1. Re:What is the magic number from? by Ancient_Hacker · · Score: 1

      Far too many postulates! And ridiculous/meaningless ones at that.

  8. Doubt it by evanh · · Score: 1

    Only when moved off DRAM will that be eliminated from the storing device.

    That said, there is probably reasonable methods of forced address order limits implemented at the controller level that can make it near impossible to compromise. This will have likely have some surprise performance hits on certain addressing patterns though.

    Roll on MRAM, I say.

  9. The current physics model is.... by Anonymous Coward · · Score: 0

    1) Standard model has particles that travel backwards in time, but no solution for the causal effects such a thing would have.
    2) Standard model has extra forces that only exist at the smallest level and disappear otherwise.
    3) It has particles as force, but no explanation as to why these particles are always there in exactly the right amount... or what they'd be like alone.
    4) It fails to tackle gravity
    5) Quantum mechanism has time travelling effects to.
    6) It has "Entanglement" which is just filtering induced correlation. I've bashed away at this many times and every experiment you do including the Delft one, simply adds a different type of filter.
    7) You have no explanation of what happens when mass turns to light and back.
    8) Astro Physics has black holes with escape velocities faster than the speed of light.... yet that requires mass, so you have a fixup to add a psuedo mass to light. To explain why massless light cannot escape a black hole!
    9) But then you have gravity propagating at C (based on observation), confusing a force and the effect of the force. You are measuring the effect not the force!
    10) And if gravity cannot escape a black hole, what is sucking in all the matter?.... so you bend space and time around the black hole to account for it.
    But then this effect 'magically' propagates at infinity, because otherwise it would not escape the black hole.
    11) You have two mechanisms that bend light, and only ever try to explain one as if its separate and different from the other (gravity and diffraction)
    12) No explanation for the speed of light
    13) You fail to address the speed of light in matter, or the difference between matter and light

    Let's be frank about physics, it's garbage. You bury your head in your equations, and hope they explain things, and you make basic logic flaws.

    This model is how it is. It's not going away. And STFU mod points, don't work.

  10. Sounds interesting but... by Hallux-F-Sinister · · Score: 3, Funny

    I’m holding out for LVPQDR5Z99 chips. The more letters and numbers it has, the more awesome it is, right? Why in 2018 are people settling for only DOUBLE data rate (DDR,) we should hold out at least for triple data rate (TDR) as a minimum!

    Honestly... are they seriously going to keep jamming more letters and numbers onto things?

    --
    Our reign has gone on long enough. Indeed. Summon the meteors.
    1. Re:Sounds interesting but... by rrohbeck · · Score: 3, Funny

      I'm rather partial to the LPXDRWTFBBQ product family.

      --
      thegodmovie.com - watch it
    2. Re:Sounds interesting but... by ShanghaiBill · · Score: 2

      people settling for only DOUBLE data rate (DDR,) we should hold out at least for triple data rate (TDR)

      That doesn't make any sense. DDR sends data on both the rising and falling edge of the clock. TDR would have no clock edge to sync the data, so it wouldn't work. If you want more data, then either increase the clock rate, or widen the bus.

      It is possible that the TDR comment was a joke, and I have been whooshed, in which case I apologize for being a humor-impaired Aspie.

    3. Re:Sounds interesting but... by Anonymous Coward · · Score: 0

      When DDR refers to the memory itself, it's shorthand for DDR SDRAM really.
      LP? that's an alternate strand of technology to "regular" DDR and obviously stands for Low Power. Hence why LPDDR means something and we have to name things to know what we're talking of.

      A pitfall of Intel laptop CPUs is they support LPDDR3 so we also care about the trailing number.

    4. Re:Sounds interesting but... by Anonymous Coward · · Score: 0

      It doesn't make sense right now, but we don't know what the future holds.

      If we for a moment assume that there is such a thing as TDR, the question would be how it works and why it would be useful.

      The clock is only used for synchronization anyway so it isn't really needed to always send the clock. USB for example relies on both units clocks being "close enough".
      For faster systems one could imagine the memory having an internal PLL and synchronizing it on the clock signal.
      This allows the memory to have a faster internal clock to work from than what is given on the wire.
      With both clock synchronized you could then do data transfer at the internal clock rate instead of the external.
      Since the data typically transitions more seldom than the clock this will reduce emitted EM.

      So, why TDR and not quad data rate then?
      Well, OK, this one would imply a 1/3 or 2/3 dutycycle of the clock and transfer in the middle of long period as well as both edges.
      Maybe the average noise from the data lanes means that there isn't really any benefit to get the data to quad transfer?
      Or maybe the clock alternates between 1/3 and 2/3 dutycycle to reduce the amplitude further?

      With an extra data line or two the data could be encoded to reduce transitions. Maybe TDR is a sweet spot.

    5. Re:Sounds interesting but... by Anonymous Coward · · Score: 0

      QDR already exists anyway.

    6. Re:Sounds interesting but... by Anonymous Coward · · Score: 0

      Yes, but actually it's not faster, it has another use. DDR is actually maxing the channel bandwidth because the memory lines are switching at the same speed as the clock. If you are using QDR just means that you've chosen a clock that runs slower than you can actually deliver, and your data lines are switching faster than it. Out of my head I think the big point of QDR was easier syncing with other devices attached to the bus because lower clocks.

    7. Re:Sounds interesting but... by Anonymous Coward · · Score: 0

      I can't tell if that's beef or pork. I'll have the beef.

    8. Re:Sounds interesting but... by thegarbz · · Score: 1

      It doesn't make sense right now, but we don't know what the future holds.

      We're talking clock synchronisation. The future definitely doesn't hold Triple anything in that regard. If you come up with a novel way of increasing data rate then for the love of god don't call it anything remotely related to DDR which has a very specific technical meaning.

      The clock is only used for synchronization anyway so it isn't really needed to always send the clock. USB for example relies on both units clocks being "close enough".

      USB devices which have two "clocks" either have problems or have synchronisation delays. The actual frequency of a USB signal differing from expected is a source of problems for many USB devices which rely on time sensitive signalling. USB may not have a clock line but that doesn't mean there isn't a specific frequency involved. That frequency is extracted from the data lines a trick that works just fine at USB speeds, and then utterly fails at the type of speeds we expect RAM chips to use.

      There are also other reasons why you would use a clock line, for example RAM uses a parallel interface, not a Universal *Serial* Bus. Making it self clocking adds a lot of expense and complexity to an otherwise incredibly simple and straight forward interface between two components.

      In any case you're completely missing the mark. The current bottleneck has nothing to do with the clock frequency. If the chips are capable of doing something faster then just raise the clock frequency, it's not close to being maxed out.

      So, why TDR and not quad data rate then?

      So let me get something clear: You don't understand what problem DDR solved. Your idea of TDR is off the mark, and QDR is something that exists (interleaving of two clock signals) and was never adopted since as I said you can just raise the clock frequency if your chips could handle it. DDR solved and electrical issue which isn't present right now. QDR may be something to look at in the future but right now it's just not necessary, and TDR is some weird solution you made up to a problem that doesn't exist.

    9. Re:Sounds interesting but... by Ancient_Hacker · · Score: 1

      .... or you send more than one bit per wire, with like four voltage levels, as already done by some pin-limited chips.

    10. Re:Sounds interesting but... by MiniMike · · Score: 1

      I think it was in jest, however possibly if you offset adjacent lines 120 degrees, like the phases in a 208 V line, you could get something resembling triple data rate, or at least something that sounded good in marketing. Of course, this triple line data rate memory would be called TLDR.

    11. Re:Sounds interesting but... by Anonymous Coward · · Score: 0

      Or going from two logic levels to more (kind of like MLC flash). Although it's probably easier to add more lines before running up against package issues.

    12. Re:Sounds interesting but... by Lost+Race · · Score: 1

      That frequency is extracted from the data lines a trick that works just fine at USB speeds, and then utterly fails at the type of speeds we expect RAM chips to use.

      LPDDR4X: 4.266 Gbps per pin.
      USB 3.0: 5 Gbps per pin.

    13. Re:Sounds interesting but... by thegarbz · · Score: 1

      USB 3.0 single device controller at one end of a SERIAL line.
      LPDDR4X: Multiple chips need to talk simultaneously sending multiple lines of data at the end of a PARALLEL line without controllers without anything in the way.

      Bandwidth of a USB3.2 device (I'll be nice since you're comparing old technology vs new): 20Gbps
      Bandwidth of a stick of LPDDR4X: 273Gbps (but the memory controller is capable of 546Gbps in dual channel)

      Now an additional reality check: USB3.2: Controllers both side of the bus. The latency for memory access application is simply horrible and is measured in the hundreds of microseconds.
      LPDDR4X: Typical latency: 10ns

      Now just to make my point I will repeat it: The frequency is extracted from the data lines, a trick that works just fine at USB speeds, and then utterly fails at the type of speeds we expect RAM chips to use.

    14. Re:Sounds interesting but... by Anonymous Coward · · Score: 0

      Personally I prefer the LPXDRZOMGWTFBBQ line, it is cooler and more hip.

  11. The consequence of this by Anonymous Coward · · Score: 0

    Suppose you build an electronic circuit exactly like today.
    Suppose you somehow were able to double the frequency of the underlying dipoles that make up that circuit.
    With Postulate G, the circuit performs *exactly* the same.

    Only the speed of light for that circuit is 2C instead of C.

    And it all works the same *but*. If this circuit (made of spinning dipoles at 2F) manages to bind to the matter around it at 1:1 (1 spin of matter to 1 spin of circuit dipole), the it will travel at C.

    Instead of being stationary, it will shoot off somewhere at the speed of light.

    This would be a tech support nightmare.... "Hello Micron.... my DRAM shot off the planet at the speed of light...."

  12. DDR4-4266 Speeds? by klingens · · Score: 1

    Why can't we get those kind of speeds for DDR4-4266 PC DIMMs? Those are almost impossible to get and then at insane prices.If RAM companies produce this kind of RAM normally, which isn't even specified by JEDEC, so they have to make it cheaper than the competition, then why are there no such DIMMs available?

    https://pcpartpicker.com/produ... shows there is basically none from Crucial/Micron. Gamers and generally Ryzen owners would pay lots of money for it. Heck, even good and expensive Crucial DDR4-3200 has Samsung chips on it cause Micron apparently can't deliver.

    1. Re:DDR4-4266 Speeds? by GabeGhearing · · Score: 4, Informative

      This is LPDDR; Intel and AMD don’t support it. Intel announced support in CannonLake that was supposed to ship in 2016... and still isn’t available to consumers.

      LPDDR4 has been standard on ARM devices(phones/tablets) for quite a few years.

      4266 is the highest rated LPDDR4 chips in the LPDDR4 spec. Even the Galaxy S9 only uses LPDDR4-3732 (1866MHz). https://www.qualcomm.com/produ...

      Maybe Apple’s new iPads use LPDDR-4266.

  13. Re: Eliminate the WEAK. by Anonymous Coward · · Score: 0

    We've identified this poster.

  14. Is that 12 Gb or GB or Gib or GiB? by nichogenius · · Score: 1

    12 Gb seems wrong. Why would DRAM chips be measured in Gb? It would make more sense to be in GiB ?

    1. Re:Is that 12 Gb or GB or Gib or GiB? by Anonymous Coward · · Score: 0

      Really, there's no good answer other than "that's the way it's done"; but one good argument is that gibibits is referring to a count of actual physical components on the chip (the individual cells), whereas gibibytes would be referring to groups of physical cells.

      But yes, you're right, the Gb the summary refers to is likely Gib (as in 2^30).

    2. Re:Is that 12 Gb or GB or Gib or GiB? by Anonymous Coward · · Score: 0

      No, what seems wrong is the '12' part. Whatever happened to powers of 2?? Shouldn't this be 16 Gb?

      And TFA quote of "Micron's 12 Gb (1.5 GB) LPDDR4X devices feature a slightly lower capacity than competing 16 Gb (2 GB) LPDDR4X offerings, but they are also cheaper to manufacture", doesn't really explain this at all.

    3. Re:Is that 12 Gb or GB or Gib or GiB? by Anonymous Coward · · Score: 0

      I think we should stop measuring it in x-bytes at all and instead measure it in active-electrons-per-minute.
      I am, however, way too wasted to even attempt to calculate that.
      I also need to poop.

  15. Thought Experiment H, Wavelength & magnetism by Anonymous Coward · · Score: 0

    Have I got your attention yet? This is where the wavelength of light comes from, and also where its magnetic field comes from.

    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 I:
    Postulate J:
    Postulate J2:
    Postulate K: How fast do forces propagate?
    Postulate L:

    -------------------
    POSTULATE H: All dipoles are equal, even red and blue light

    Dipoles's spin is an in-place oscillation.

    Each pole moves towards the center and back out to the extreme. Tracing out a flattened wave along the diameter.

    It's easier to understand it in terms of energy. At the extremes they are all potential energy across their electric field.
    At the center they are all kinetic energy (deriving from velocity over electric force).

    They have no mass, there is no inertia at this level. Just energy, and electric force.
    So they simply flip between kinetic, potential, kinetic potential, pulsating rather than spinning.

    If you applied a force to an *undisturbed* dipole, the oscillation turns into a spin.

    Imagine a single isolated dipole oscillating up and down. Push it left, and you've spread out the up-down movement into a ellipse.
    Stop pushing and it returns to an up- down oscillation along a line.

    ---------------
    All dipoles are the same, same frequency, same magnitude, all coherent.
    They all have exactly the same energy. So how does wavelength of light work?

    Red light has a longer wavelength than blue light.
    Red light has less energy than blue light. It is less energy dense.

    Dipole R is red light, it has wavelength WR
    Dipole B is blue light, it has wavelength WB
    Dipole M is matter, it has a wavelength of W

    The wave of red light takes WR/W dipolar oscillation to trace out.
    The wave of blue takes WB/W dipole oscillations to trace out
    WR/W > WB/W, i.e. it takes more spins to trace out red's wave than blue's wave.

    Conversely the density of energy in red light is less.
    W/WR W / WB

    Per spin (the unit of time Postulate G), red light provides less energy than blue light!

    If a dipole is spinning at the same frequency with the same magnitude, how can it take longer to trace out a wave?
    It's really really simple.

    For undisturbed matter the dipole oscillation moves up and down in a line.
    For light it rotates about an axis normal to the direction of travel and normal to the oscillation.

    Note that light must have a second field if it spins in a plane like this.... i.e. it's magnetism comes from the circular spin.

    Postulate A: https://science.slashdot.org/comments.pl?sid=12866516&cid=57598384
    Postulate B: https://science.slashdot.org/comments.pl?sid=12869268&cid=57599376
    Postulate C: https://news.slashdot.org/comments.pl?sid=12870718&cid=57604074
    Postulate D: https://hardware.slashdot.org/comments.pl?sid=12877158&cid=57615478
    Postulate E & E2: https://hardware.slashdot.org/comments.pl?sid=12877158&cid=57615516
    Postulate F: https://hardware.slashdot.org/comments.pl?sid=12877158&cid=57615700
    Postulate G: https://hardware.slashdot.org/comments.pl?sid=12877158&cid=57615906

  16. We've come a long way by Anonymous Coward · · Score: 0

    So we've covered, mass (binding potential, not real), force (all derive from electric), time (periodicity of spin), speed of light (wave length of spin per spin), why the effects of electric force take time to propagate but the electric force is instant.

    You've got a simple understanding of how light and matter are the same thing.

    We've got every dipole identical, all the same energy, all the spin dipole size.

    Despite this, we have waves that travel at C, and have different energys and different wavelengths DESPITE BEING TRACED OUT BY THE SAME DIPOLES WITH THE SAME ENERGIES.

    And I've explained in a way that a 14yo can understand.

    1. Re:We've come a long way by Anonymous Coward · · Score: 0

      I'm 14 but still can't understand you. May if you really committed and used all caps I would.

    2. Re:We've come a long way by Anonymous Coward · · Score: 0

      If this series doesn't end with blueprints for a warp drive or transporter disk, I'm going to be very disappointed.

  17. It all started with by Anonymous Coward · · Score: 0

    IGBT

    Then the innvation-cravers who worship money and material thought they needed something like this too. Now we have LQBTABCDEFXZ. And people who identify as "diaper fetishist".

  18. Re:Thought Experiment by Anonymous Coward · · Score: 0

    This might be useful to you.

    It was not.

    It explains how light matter and time work.

    It does not.

    And its simple enough to explain in reasoned postulates.

    It is simple enough for you to understand. The actual universe is not limited by your level of understanding.

    Conduct a repeatable experiment that supports your theories and violates the Standard Model. Until then this is garbage.

  19. 12? Where did the other 4 go? by Anonymous Coward · · Score: 0

    12 is a strange number. I'm used to seeing ram come in some power of 2, i.e., 4, 8, 16, etc.

    Do these chips need a special memory controller or are memory controllers configurable to handle any chip size?