4 ms·

Models this small and this capable bode really well for the usefulness of a PC like the RTX Spark that Nvidia/Microsoft announced this week. 128GB of unified me

Models this small and this capable bode really well for the usefulness of a PC like the RTX Spark that Nvidia/Microsoft announced this week. 128GB of unified memory will likely be more than sufficient for effective local agentic coding, even if SOTA cloud models will still be even better. Up until this point, I've found the cost/value to unequivocally favor using a cloud subscription, but I would be lying if I didn't worry that one day OpenAI is going to increase the price for my subscription by 5-10x. I rely on these tools enough that if there is a real viable local option, I'm going to take it.

pseudollm 4mo ago
> usefulness of the RTX Spark Not really. There's a reason the announcement didn't include ANY benchmark (!) and didn't mention EXACTLY what is the memory bandwidth. It's going to be dog-slow unusable for large models, as tok/sec is basically bandwidth divided by active weights. Rumoured 300GB/s / 30GB active weights (decent model) = 10 tokens per second, which is really slow
- SwellJoe 4mo ago
  Yep, I have a Strix Halo and while it can run models bigger than Qwen 3.6 27b, it's not usable interactively when you do. ds4 patched for ROCm works, but at such a slow speed, it's not usable for coding agents. The Nvidia boxes have only slightly more memory bandwidth, so I wouldn't expect them to be notably faster. At least not enough to make it useful interactively at that scale.
  - zozbot234 4mo ago
    Why does everyone expect interactivity from local AI? It's not the best use of the hardware, especially not miniPC hardware. Long-term batched inference with larger and more capable models is much more feasible AIUI.
    - SwellJoe 4mo ago
      Even batched it's uncomfortably slow. I started to benchmark ds4 with my security vulnerability benchmark (after Qwen 3.6 dense and MoE and a bunch of cloud models), but it was going to tie up the Strix Halo for more than a day, so I decided not to run it as it would prevent me from doing other stuff with it during that time. Even batched usage needs to be fast enough to deliver results in a reasonable time. Overnight runs are useful, 24 hour runs are...less so. Anyway, most of the time people are talking about interactive use, and there's currently an upper bound on how large a model can be for local hosting on a reasonable budget (i.e. not a crazy amount more expensive than what a high end developer desktop or laptop costs). The sweet spot is probably currently the big Qwen 3.6 or Gemma 4 models, which are in the ~60GB range for 8-bit quantization plus a large context.
      - hedgehog 4mo ago
        The 6-bit versions + 8-bit KV cache seems to save a good bit of memory without a significant loss of quality. The Qwen 35B is pretty fast in my testing, but MiniMax M2.7 230B is in some ways faster (way fewer tokens to arrive at an answer) even though it is much larger.
        SwellJoe 4mo ago
        Qwen 3.6 35B-A3B with MTP at 8 bits is blazing fast, something like 50-60 tokens per second. That's plenty fast for interactive use, so I haven't tried lower bits. Unfortunately the MoE is notably dumber than the dense model (for the case I have data about...I've been benchmarking models for security vulnerability scanning, and 27B is notably better on hard bugs). The dense model is almost usable, but feels really sluggish, even with MTP. I think it's about 12-15 tokens/second on the Strix Halo. Slow enough to where I'd rather pay to use a cloud model. I might try the 6-bit version of the dense model to see how it behaves, though. Maybe it'll retain its bug hunting abilities while making it fast enough to use interactively and not take all day for benchmark runs.
        hedgehog 4mo ago
zozbot234 4mo ago
RTX Spark is pretty much the DGX Spark in a laptop form factor, plus some lower-performing chips in the same series to be released later according to rumors. We know quite well how the top-of-the-line chip performs: it's very interesting for some application areas, less so for others.
dirkg 4mo ago
The RTX/DGX Spark, Mac Ultras with 128GB unified ram are all ~$5k. Its still an expensive toy for rich people, it might as well be an H100 for 99.9% of the population (not devs with high paying jobs, of course). the value of local models is allowing normal people to access AI without needing to subscribe to cloud services. this is esp imp for the rest of the world where even a 12GB gpu is extremely expensive. there is no real viable local option that will come even close to Sonnet/Gemini Flash or the cheaper chinese models. Even if your pc costs <$2k you are never going to recoup the hw costs, and the results will be far worse.
- green7ea 4mo ago
  I'm using a Strix Halo laptop (~3k, 64GiB) and with Gemma 4 and Qwen 3.6, both at 8 bits, I'm seeing very impressive results. As a work tool, this is reasonably priced. You can save a bit of money by opting for a non-laptop form factor.
- organsnyder 4mo ago
  My Framework Desktop with 128GB was about half that. I did luck out by buying before RAM prices went crazy, though. I'm looking forward to the fallout when the data center bubble bursts. There's a good possibility we'll see a glut of hardware, either on the used market or from manufacturers that no longer have massive orders from OpenAI and the like.

pseudollm 4mo ago

> usefulness of the RTX Spark Not really. There's a reason the announcement didn't include ANY benchmark (!) and didn't mention EXACTLY what is the memory bandwidth. It's going to be dog-slow unusable for large models, as tok/sec is basically bandwidth divided by active weights. Rumoured 300GB/s / 30GB active weights (decent model) = 10 tokens per second, which is really slow

SwellJoe 4mo ago
Yep, I have a Strix Halo and while it can run models bigger than Qwen 3.6 27b, it's not usable interactively when you do. ds4 patched for ROCm works, but at such a slow speed, it's not usable for coding agents. The Nvidia boxes have only slightly more memory bandwidth, so I wouldn't expect them to be notably faster. At least not enough to make it useful interactively at that scale.
- zozbot234 4mo ago
  Why does everyone expect interactivity from local AI? It's not the best use of the hardware, especially not miniPC hardware. Long-term batched inference with larger and more capable models is much more feasible AIUI.
  - SwellJoe 4mo ago
    Even batched it's uncomfortably slow. I started to benchmark ds4 with my security vulnerability benchmark (after Qwen 3.6 dense and MoE and a bunch of cloud models), but it was going to tie up the Strix Halo for more than a day, so I decided not to run it as it would prevent me from doing other stuff with it during that time. Even batched usage needs to be fast enough to deliver results in a reasonable time. Overnight runs are useful, 24 hour runs are...less so. Anyway, most of the time people are talking about interactive use, and there's currently an upper bound on how large a model can be for local hosting on a reasonable budget (i.e. not a crazy amount more expensive than what a high end developer desktop or laptop costs). The sweet spot is probably currently the big Qwen 3.6 or Gemma 4 models, which are in the ~60GB range for 8-bit quantization plus a large context.
    - hedgehog 4mo ago
      The 6-bit versions + 8-bit KV cache seems to save a good bit of memory without a significant loss of quality. The Qwen 35B is pretty fast in my testing, but MiniMax M2.7 230B is in some ways faster (way fewer tokens to arrive at an answer) even though it is much larger.
      - SwellJoe 4mo ago
        Qwen 3.6 35B-A3B with MTP at 8 bits is blazing fast, something like 50-60 tokens per second. That's plenty fast for interactive use, so I haven't tried lower bits. Unfortunately the MoE is notably dumber than the dense model (for the case I have data about...I've been benchmarking models for security vulnerability scanning, and 27B is notably better on hard bugs). The dense model is almost usable, but feels really sluggish, even with MTP. I think it's about 12-15 tokens/second on the Strix Halo. Slow enough to where I'd rather pay to use a cloud model. I might try the 6-bit version of the dense model to see how it behaves, though. Maybe it'll retain its bug hunting abilities while making it fast enough to use interactively and not take all day for benchmark runs.
        hedgehog 4mo ago

zozbot234 4mo ago

RTX Spark is pretty much the DGX Spark in a laptop form factor, plus some lower-performing chips in the same series to be released later according to rumors. We know quite well how the top-of-the-line chip performs: it's very interesting for some application areas, less so for others.

dirkg 4mo ago

The RTX/DGX Spark, Mac Ultras with 128GB unified ram are all ~$5k. Its still an expensive toy for rich people, it might as well be an H100 for 99.9% of the population (not devs with high paying jobs, of course). the value of local models is allowing normal people to access AI without needing to subscribe to cloud services. this is esp imp for the rest of the world where even a 12GB gpu is extremely expensive. there is no real viable local option that will come even close to Sonnet/Gemini Flash or the cheaper chinese models. Even if your pc costs <$2k you are never going to recoup the hw costs, and the results will be far worse.

green7ea 4mo ago
I'm using a Strix Halo laptop (~3k, 64GiB) and with Gemma 4 and Qwen 3.6, both at 8 bits, I'm seeing very impressive results. As a work tool, this is reasonably priced. You can save a bit of money by opting for a non-laptop form factor.
organsnyder 4mo ago
My Framework Desktop with 128GB was about half that. I did luck out by buying before RAM prices went crazy, though. I'm looking forward to the fallout when the data center bubble bursts. There's a good possibility we'll see a glut of hardware, either on the used market or from manufacturers that no longer have massive orders from OpenAI and the like.