GlobalFoundries unveils 7nm-class FDX Fusion node — FD-SOI node doubles 22FDX transistor density, comes 7 years after firm abandoned FinFET

GlobalFoundries
(Image credit: GlobalFoundries)
8

When GlobalFoundries abandoned development of leading-edge nodes in 2018, it also canceled the development of its 7nm-class FinFET fabrication technology that it co-developed with IBM. Almost exactly eight years later, GlobalFoundries is back with a 7nm-class manufacturing process, only this time the node relies on planar transistors and fully-depleted silicon-on-insulator (FD-SOI). It is aimed at automotive, edge AI, industrial, robotics, sensors, and other devices that prioritize low power consumption, high integration, and cost efficiency over maximum performance.

"We are delighted that GlobalFoundries has announced its next-generation 'single-digit FD-SOI; node, building on technology that CEA-Leti has been advancing for decades," said Sébastien Dauvé, chief executive of CEA Leti. "Through the FAMES pilot line, CEA-Leti continues to shape the future European roadmap for FD-SOI and works every day to promote its adoption in a wide range of applications."

FDX Fusion enters the scene

Latest Videos FromTom's Hardware
Anton Shilov
Contributing Writer

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

  • bit_user
    It's nice to see they're getting the ball rolling on process improvements, at least. I hope it won't be nearly as long before their next node introduction.
    Reply
  • thestryker
    As far as I'm aware they haven't acquired any EUV machines so it's probably a safe bet that the node won't match existing "7nm" nodes. I believe all of the nodes on the market with similar application are based on 20nm or higher original nodes though so there's definitely a lot of room for improvement.

    Given that their aborted "12" node was based on their 14nm process it has always made me wonder if there was something simply unsuitable which would require a purpose built node. It definitely sounds like this forthcoming node isn't going to be built directly off of anything existing.
    Reply
  • SonoraTechnical
    Are they doing any of this in their Malta, NY facility? That one has been stuck on older nodes for some time.
    Reply
  • palladin9479
    Now if only they could make DRAM..
    Reply
  • usertests
    palladin9479 said:
    Now if only they could make DRAM..
    Time to use SRAM as DRAM
    Reply
  • bit_user
    usertests said:
    Time to use SRAM as DRAM
    Although I'm sure you're just being a bit silly, it's a good chance to review their relative densities.

    According to a quick search, the 64 MB 3D cache die is 36 mm^2 on Zen 5; 41 mm^2 on Zen 4. So, that works out to about 1.78 MB/mm^2.

    Micron's 32 Gb DDR5 die is reportedly 71 mm^2. So that works out to 57.7 MB/mm^2. That's roughly 32x the density of the above SRAM.

    So, um... no. We won't be doing that any time soon.

    (source: Google search. The results seem plausible enough to me, within the margins required for this back-of-envelope calculation, that I didn't bother to follow the links and confirm. You're welcome to, if you wish.)
    Reply
  • usertests
    bit_user said:
    Micron's 32 Gb DDR5 die is reportedly 71 mm^2. So that works out to 57.7 MB/mm^2. That's roughly 32x the density of the above SRAM.
    All my homies need 32-layer 3D SRAM, pronto!
    Reply
  • SemiChemE
    thestryker said:
    As far as I'm aware they haven't acquired any EUV machines so it's probably a safe bet that the node won't match existing "7nm" nodes. I believe all of the nodes on the market with similar application are based on 20nm or higher original nodes though so there's definitely a lot of room for improvement.

    Given that their aborted "12" node was based on their 14nm process it has always made me wonder if there was something simply unsuitable which would require a purpose built node. It definitely sounds like this forthcoming node isn't going to be built directly off of anything existing.
    GF's 12nm Finfet node(12LP+) based on their 14nm node has been in production since at least 2020, if you owned a Samsung TV in the early 2020's it probably included one of these chips.

    Don't confuse it with 12FDX, which was supposed to be the successor to 22FDX (FDSOI = Fully Depleted SOI), which instead of using a "fin" uses a planar thin silicon channel on an insulator. It can't really compete with Finfet on performance, but it's much cheaper, can be configured to operate in ultra low power modes, is radiation hard, has huge advantages for RF, and due to its relative simplicity can readily be integrated with MRAM or other forms of NVM. It's also easier to qualify for high reliability applications like automotive. It really hits the sweet spot for a lot of IOT applications. But, since it requires very specialized design libraries, the challenge for GF has always been about timing the release so as not to cannibalize the 22FDX business. FDX Fusion appears to be a rebranding of 12FDX, but perhaps with some new performance enhancements to bring it closer to 7nm Finfet performance (and qualify for EU research funds targeted at sub-10nm nodes).
    Reply