A practical guide to running 8x RTX PRO 6000's

(gpupartner.com)

17 points | by Retell15 2 days ago

11 comments

  • srcreigh 7 minutes ago
    Makes you realize how insane the M5 Ultra Mac Studio is. 1.2GB/s bandwidth 512GB memory. Its rated max power draw is just 480W. And it also has amazing M-series CPUs. It costs less than just one of these GPUs which each take 700W to run.
  • schaefer 52 minutes ago
    > We currently have 14x nodes of CG480-S6053 ready to ship.

    Oh, okay, so this is an ad.

    I do still think it's well written and interesting... But if anything, it's just making me more curious about the newest generation of M5 Ultra. (and less and less interested in PCI-E Gen 5 anything)

    • baron3dl 49 minutes ago
      You must be the only one 'round here without a stack of RTX PRO 6000s, 8 high, that you're not sure how to use.
      • schaefer 44 minutes ago
        what I have is a DGX Spark to play and learn on, and an OpenRouter account for actual work.

        I mostly rent my tokens.

        • baron3dl 35 minutes ago
          that was an attempt at monocle and top hat irony.

          edit: how do you feel about paying 5% on every token's cost to stripe? it made me cancel, and sign up directly with a few providers.

          • schaefer 16 minutes ago
            > attempt at ... irony

            For the record, I did laugh, I just didn't type it out.

            > how do you feel about paying 5% on every token's cost to stripe?

            Life is compromise. In a perfect world, I'd love to buy a linux native system comparable to the Mac Studio that could run something in the ballpark of Deepseek V4 flash "fast" at home.

            Not only does that not exist, but even if it did I couldn't justify the cost. Between work and home my annual token spend is maybe 3k.

            So... paying a 5% fee on a service that allows me to not shell out 10k on hardware at home still seems like a pretty sweet deal.

            --

            Plus, I still am learning a ton and playing lots. It's hard to imagine anything beating openrouter for that. So much exposure to all the latest LLMs...

  • jimmoores 24 minutes ago
    These people have zero idea what they're doing. Not a single mention of pipeline parallelism that would actually make the setup useful to run a big model.
    • itkovian_ 6 minutes ago
      I can’t stand it. Very engineering-y over specified formal language around a complete lack of core understanding. Is damaging other people read this and try to learn things from it.
    • schaefer 12 minutes ago
      can you point to a write up that discusses what you're talking about?

      because I would read it.

    • kelmoran 19 minutes ago
      I feel like you would want to run 8 smaller models separately for quantity of raw output. 1 big model is slow and isnt guaranteed to make no mistakes.
  • RachelF 48 minutes ago
    For those who can't afford RTX 6000's you can unlock around 20% increased card to card speed on consumer GPUs using this library:

    https://github.com/aikitoria/open-gpu-kernel-modules

    The hardware supports it, but Nvidia disabled it if the driver detects cheaper cards.

  • kmike84 3 minutes ago
    Pass. When articles keep mentioning models like DeepSeek R1, or Llama 3.1, or Qwen3 32B, it is a pretty robust indicator of AI slop. LLMs love to suggest DeepSeek R1, etc. - training data cut-off?

    No person with real practical experience and real use cases will be using these ancient models as examples, when talking about local LLMs.

  • xyst 13 minutes ago
    SLI is relevant again
  • CamperBob2 26 minutes ago
    4x RTX 6000 Blackwell cards is a good place to be if you can't swing 8 of them, or if you don't have the power or cooling to run that many. A system based on 4x RTX6K can run GLM 5.3 at NVFP4 precision [1] from a US-standard 120V 20A circuit when derated to 300W, and give you a better pelican than Fable 5.1 [2]. What's not to like?

    (Edit: I'm mistaken here, the pelican didn't come from Flash on 4 cards but from the full GLM 5.3 model on 8. But the Flash model is still crazy good for its size.)

    1: https://huggingface.co/local-inference-lab/GLM-5.3-NVFP4

    2: https://crimson-jeri-74.tiiny.site/

    • nojs 3 minutes ago
      > A system based on 4x RTX6K can run GLM 5.3 at NVFP4 precision

      It actually runs fine at FP8 on this hardware too, with the full 1M context.

  • robotnikman 29 minutes ago
    Now if only I could afford 8 RTX PRO 6000's
    • CamperBob2 23 minutes ago
      Start with 4, see my other comment. The recent GLM, Qwen and DeepSeek releases are amazingly promising.
      • jplusequalt 5 minutes ago
        >Start with 4

        Sure, let me just buy $60,000 worth of GPUs to run a *quantized non-frontier model*.

        For that price you could:

        - put a down payment on a home in a large % of the US

        - buy a brand new car in cash (possibly two!)

        - take a long sabbatical and travel the world

        - pay all 4 years or your child's college tuition

      • estebarb 16 minutes ago
        Now, if I could afford 4...
  • christkv 50 minutes ago
    600W * 8 just for the GPUs when maxed out (besides the cost). Def nothing for my home lab.
    • touisteur 5 minutes ago
      I'm curious whether actual inference workloads actually push to 600W (and not 350W) and what the last 250W get you. Rare is the (generic gpu) workload where I get >5%, some rare light inference benchmarks up to 10%...
  • varispeed 53 minutes ago
    > (~9.2 million tokens node-wide at 4k context).

    stopped reading after that. What 4k context would be usable for?

    • monster_truck 9 minutes ago
      So much more than you'd realize!

      That's a solid 2200 words to spend on operating parameters and conveying state, leaving a generous 700 word window for them to decide and respond in.

      When the bonsai/prism 1bit models dropped and I saw how many prompts a minute I could get from a dusty m2 mini I started hooking it up to all sorts of shit, like a traffic simulator that translates the car state/surroundings/immediate goal into text, it responds with a seqeunce of actions defined in the system prompt, which then get translated back into NPC input.

      What I was hoping for here was that it would result in fucking chaos, all sorts of stupid decisions and epic car accidents. I cannot overstate my disappointment (and terror) when they were perfectly reasonable, safe drivers. I had to cut the tire grip by 75% without telling them and make them control twice as many cars to delay their ability to respond before I saw anything resembling an enjoyable traffic accident.

    • erdaltoprak 17 minutes ago
      Not long-horizon coding but for a lot of other things like batch processes with structured outputs, quick checks/fixes, making sense of unstructured data etc..
  • inventor7777 19 minutes ago
    This will really help when my 8 RTX PRO 6000s ship. /s