• lime!@feddit.nu
    link
    fedilink
    arrow-up
    6
    arrow-down
    1
    ·
    24 hours ago

    a 128GB framework desktop could do that job. it’s increased a bit in price since i last looked at it but €4500 isn’t that much for a company.

      • lime!@feddit.nu
        link
        fedilink
        arrow-up
        4
        arrow-down
        1
        ·
        23 hours ago

        i’m running moderately quantized models on 24GB VRAM and getting like 30-40 tokens a second. add a zero to the price and it’s still not a lot for a company.

        • theunknownmuncher@lemmy.world
          link
          fedilink
          arrow-up
          4
          arrow-down
          1
          ·
          edit-2
          23 hours ago

          Sure, but you’re running a very small model compared to what we are talking about.

          GLM-5.1 is over 200GB even when quantizied to 1-bit. Kimi K2.6 is even bigger. A framework desktop cannot run either of these. Qwen3.6 is significantly smaller and the model weights could fit, but consider the KV-cache you’d need for all of the company’s users, and the throughput required to serve them all.

          You’re right that it is within reach for a company but framework desktop makes zero sense for this

          • lime!@feddit.nu
            link
            fedilink
            arrow-up
            2
            ·
            23 hours ago

            isn’t qwen like 40-50GB? that could work i think. performance is okay even quantised down to 10.

            • Evotech@lemmy.world
              link
              fedilink
              arrow-up
              1
              ·
              15 hours ago

              And then add 200k context on top

              And then add hundred of users needing to do things in paralell

              • boonhet@sopuli.xyz
                link
                fedilink
                arrow-up
                1
                ·
                13 hours ago

                If it’s a large enough company to have hundreds of users, it can afford several beefy machines tbh

                • Evotech@lemmy.world
                  link
                  fedilink
                  arrow-up
                  1
                  ·
                  12 hours ago

                  It’s a capex and that type of hardware needs to be replaced every 3 years minimum and you need people to set it up and maintain a cluster. And it’s not straight forward.

                  You are never going to get that approved without a serious business case.

                  Claude on the other end is a opex and much easier to just try out and then build a solution on it

                  Not saying it doesn’t happen but it’s not as easy as people make it sound like

              • lime!@feddit.nu
                link
                fedilink
                arrow-up
                1
                ·
                14 hours ago

                nobody said anything about it being a large company :P

                anyway, seems the framework is hampered by a slow gpu so the memory issues are apparently moot.

            • Jiral@lemmy.org
              link
              fedilink
              arrow-up
              1
              arrow-down
              1
              ·
              21 hours ago

              At Q8 it is around 35-40GB I think + memory for required context.

              I have a Framework desktop. It gets you you around 6t/s. Not suitable for professional use but for personal use I think it is fine. I do prefer Gemma 4 though, but that comes with similar reqirements.

              • lime!@feddit.nu
                link
                fedilink
                arrow-up
                1
                ·
                21 hours ago

                huh, i thought that ryzen ai thing would perform better than that. my 7900xtx regularly gets 30+tps with qwen, up to hundreds with more compressed models.

                • Jiral@lemmy.org
                  link
                  fedilink
                  arrow-up
                  2
                  ·
                  21 hours ago

                  My system runs at 100W TDP though. That is maybe 140W at the power outlet, incl. monitor and everything.

                  This is also the dense 27B model at Q8. But yeah, it is not terribly fast. I think the best use case is on MoE models. GPT-OSS-120B runs on it for example and at 50T/s speed is not a n issue anymore either. (I could get it to run even on just 64GB but the new llama.cpp might need a tiny bit more memory which pushed it just across the limit. yeah I know, for seriously using it you’d need the 128GB version)

                  • lime!@feddit.nu
                    link
                    fedilink
                    arrow-up
                    2
                    ·
                    21 hours ago

                    that’s fair, i’m at like 7x the power. the gpu alone easily pulls 350-400W and the rest of the system isn’t exactly running lean either.

                    …man now i really want more vram.