• raicon@lemmy.world
    link
    fedilink
    English
    arrow-up
    12
    ·
    3 days ago

    Plan B in case openai breaks this year? Come back to gamers begging on their knees

    • Rekall Incorporated@piefed.socialOPM
      link
      fedilink
      English
      arrow-up
      3
      ·
      3 days ago

      GeForce “Rubin” will be built on some variant of the current TSMC 3 nm FinFET node. They need not be the same N3 node that’s in use by Apple, Intel, and others; and NVIDIA might collaborate with TSMC on an exclusive variant just the way it created the NVIDIA 4N node, derived from TSMC N5.

      I will speculate power efficiency will be at worst roughly equal to the 5000 series on a SKU class basis (5060 to 6060 and so on), most likely we will see efficiency gains at the SKU class level.

      There was a big jump between the 3000 and 4000 series as Nvidia went from Samsung 8 NM to TSMC 4N for the 4000 series.

      Both the 4000 and 5000 series were on the same TSMC 4N node.

      For the 6000 series we are expecting some sort of new TSMC node (a variant of 3NM), so I would expect efficiency (at the SKU class level) to improve or at worst stay the same.

  • 9point6@lemmy.world
    link
    fedilink
    English
    arrow-up
    3
    ·
    3 days ago

    Wait 1/3 of the raster performance seems like a massive downgrade, right? Or am I reading that wrong

    • Rekall Incorporated@piefed.socialOPM
      link
      fedilink
      English
      arrow-up
      5
      ·
      3 days ago

      I am assuming they are saying raster performance will be 30% to 35% higher than the 5000 series.

      There is no way raster would be 35% of the performance of the 5000 series.