back to article Toshiba extends enormo flash fab, hopes to bust NAND capacity barrier

Toshiba has said it will expand its Fab 5 NAND fabrication plant at Yokkaichi Operations in Mie, Japan, following a SanDisk announcement about the same effort. The massive Yokkaichi fab plant apparently has an underground monorail to shuttle material from one building to another, so the expansion would mean adding at least one …


This topic is closed for new posts.
  1. ScissorHands

    Stacking dies isn't a silver bullet

    Process shrinking allows more cells on the same die area, increasing productivity. It also means £/GB drops following Moore's Hunch (TM) and approaches spinned-rust levels.

    Unfortunately Flash is hitting a wall where it can't be shrunk any further...

    Die stacking only increases the density (how much storage you can cram into a 2.5" SSD), while doing almost nothing about the £/GB or power consumption.

    1. Nigel 11

      Re: Stacking dies isn't a silver bullet

      I don't think power consumption is too much of a problem for the forseeable future. An SSD uses something like one watt spread out over maybe 40 square centimeters. (And note, some of that is the controller not the flash chips). A CPU or other highly active logic chip uses maybe 100 watts originating in maybe one square centimeter, but spread out over a heatsink footprint of maybe the same 40 square centimeters. So provided there's good thermal connectivity between stacked chips, one could stack a hundred of them (with a decent heatsink on top) without exceeding the thermal loadings that are commonplace in desktop PC CPUs and high-end graphics cards. (I doubt we'll be seeing 100-deep stacks any time soon, but thermal loading isn't the reason).

      Price/Gbyte? Well, it'll be interesting to see just how far and how fast economies of (vast!) scale can drive down the price of manufacturing flash chips.

  2. John Smith 19 Gold badge

    Now if you could build a solid stage magnetic disk system.

    Just saying.

    Some attempts were made in the 1950s with a sort of inverse magnetic ring system using holes in a NiFe alloy sheet to localize the fields and electroplated copper (newly developed for the printed circuit board industry).

    Time to have another go?

    1. Greg D

      Re: Now if you could build a solid stage magnetic disk system.

      The problem with magnetic storage is the read/write speeds and seek times. Flash solved a huge latency issue when it comes to read/write cycles, and all but removed the disk bottleneck on modern computers. Magnetic storage has no future in this regard. It's gone as far as it can go methinks.

      1. Nigel 11

        Re: Now if you could build a solid stage magnetic disk system.

        I think magnetic storage will hold on at the large data end. When one combines bit-patterned media and heat-assisted recording (BPM and HAMR) one can anticipate drive capacites measured in tens of Terabytes. This is technology that's already working in the lab. I should think that the question at present, is whether anyone will ever want single drives that size? Having Terabyte flash caches up front may well drive demand.

This topic is closed for new posts.

Other stories you might like