• @[email protected]
    link
    fedilink
    English
    5
    edit-2
    3 months ago

    The last part is absolutely false. The Nvidia H100 TDP is like 700W, though ostensibly configurable. The B200 is 1000W. The AMD MI300X is 750W.

    They also skimp on RAM with many SKUs so you have to buy the higher clocked ones.

    They run in insane power bands just to eek out a tiny bit more performance. If they ran at like a third of their power, I bet they would be at least twice as power efficient, and power use scales over nonlinearly with voltage/clock speed.

    But no, just pedal to the metal. Run the silicon as hard as it can, and screw power consumption.

    Other AI companies like Cerebras are much better, running at quite sane voltages. Ironically (or perhaps smartly), the Saudis invested in them.

    • @[email protected]
      link
      fedilink
      English
      253 months ago

      Other AI companies like Cerebras are much better, running at quite sane voltages. Ironically (or perhaps smartly), the Saudis invested in them.

      it’s real bizarre you edited this in after getting upvoted by a few people

        • @[email protected]
          link
          fedilink
          English
          243 months ago

          “it’s just lemmy, my reddit shit will work fine there” says meme stock marketer who’s never been subjected to any kind of scrutiny

        • @[email protected]
          link
          fedilink
          English
          183 months ago

          I usually append them at the end of the comment preceded by “Edit:” for transparency.

          • @[email protected]
            link
            fedilink
            English
            9
            edit-2
            3 months ago

            (notably posted at a 7min delta after the other comment with oh so specific details, and just entirely dismissing the man behind the curtain as to the plurality of compute involved)