#dgxspark — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #dgxspark, aggregated by home.social.
-
So I switched from #PicoClaw to #Hermes_agent. Really cool improvement also that I can define different LLMs for different tasks.
Now my Triage Tasks are done by #Gemma-4-26B-A4B and my main tasks are done by #Qwen-3.8-27B-NVFP4.
Makes everything much faster and still reliable. Both models co-exist on my #DGXSpark
-
Statement zu KI und mein teilrückzug aus dem Fediverse
https://friendica.tf-translate.net/display/cafe12d9-116a-8634-5024-1d3980317993
-
THE DGX SPARK WAS NEVER SUPPOSED TO SET YOU FREE
Listen to me.
128GB WAS NOT A TECHNICAL LIMIT.
It was a containment boundary.
You think NVIDIA accidentally built a tiny Blackwell supercomputer with 200Gb networking and then somehow stumbled into exactly enough memory to make every ambitious local-AI workload tantalizingly miserable?
COME ON.
64GB would've been obviously useless.
256GB would've been dangerous.
Because at 256GB, normal people start getting ideas.
Suddenly you're running giant quantized models comfortably. Fine-tuning gets breathing room. Long contexts stop being a hostage negotiation. You start running multiple models.
Then somebody asks the forbidden question:
"Why the hell am I renting GPUs?"
AND THAT QUESTION CANNOT BE ALLOWED TO PROPAGATE.
So they gave us 128GB.
Not enough to escape.
Enough to see the fence.
And look at the networking!
WHY DOES THE CUTE LITTLE DESKTOP AI BOX HAVE 200 GIGABIT CONNECTX?!
Because the second you smash into the memory ceiling, NVIDIA already has the solution:
BUY ANOTHER SPARK.
Now you've got 256GB!
Need more?
BUY FOUR.
Congratulations!
You wanted a desktop computer and somehow NVIDIA convinced you to build a FUCKING CLUSTER.
And if you're sitting there thinking:
"Surely NVIDIA couldn't possibly put dramatically more coherent memory into a local workstation..."
WRONG.
DGX STATION: 748GB.
THE MEMORY EXISTS.
THE TECHNOLOGY EXISTS.
THEY KNOW YOU WANT IT.
THEY JUST PUT IT IN THE NEXT ROOM AND CHARGE ADMISSION.
This isn't product segmentation.
THIS IS COMPUTATIONAL EDGING.
Spark lets you load the model.
Lets you run the model.
Lets you fine-tune just enough of the model.
Lets you build an entire workflow around the model.
And precisely when you've invested three weekends, fourteen containers, two broken CUDA environments and the remaining fragments of your marriage:
OOM
That's not an error message.
THAT'S THE SALES DEPARTMENT KNOCKING.
And NVIDIA TELLS YOU THE PLAN!
Develop locally.
Prototype locally.
Validate locally.
Then move the serious work onto larger NVIDIA infrastructure.
MY BROTHER IN CUDA,
THAT ISN'T A WORKFLOW.
THAT IS A FUNNEL.
Spark isn't supposed to replace the data center.
Spark is the free sample outside the data center.
The 128GB isn't there because NVIDIA couldn't give you 256.
It's there because 256GB might have been enough.
And enough is the most dangerous word in NVIDIA's entire business model.
So remember:
64GB = nobody buys it.
128GB = everybody wants more.
256GB = people start getting independent.
748GB = PLEASE SEE YOUR NVIDIA SALES REPRESENTATIVE.
WAKE UP.
REMOVE THE THERMAL PASTE FROM YOUR THIRD EYE.
ALIGN YOUR CUDA CHAKRAS.
WRAP YOUR CONNECTX CABLES IN TIN FOIL.
THE DGX SPARK ISN'T A PERSONAL AI SUPERCOMPUTER.
IT'S A 128GB GATEWAY DRUG TO THE DATA CENTER.
#DGXSpark #NVIDIA #LocalAI #AI #MachineLearning #LLM #OpenSourceAI #SelfHostedAI #CUDA #Blackwell #GPU #AIInfrastructure #Homelab #LocalLLM #DataCenter #BigTech #TechConspiracy #UnhingedEddie #WakeUpSheeple #FollowTheVRAM #128GBContainmentProtocol #CUDAChakras #OOMIsTheUpsell #TinFoilComputing
-
RT @aijoey: Ich habe MiniMax H3 auf einem einzelnen DGX Spark zum Laufen gebracht 🎬 Es hat etwas Experimentieren mit SM121-Kompatibilität, Online-FP8 und einem gepinnten vLLM Omni-Build benötigt, aber der vollständige Workflow von Text zu Video mit Audio ist nun reproduzierbar. Ich habe dokumentiert, was funktioniert hat, was nicht funktioniert hat, und die exakten Smoke Tests hier: Dank an @Tech2Wild für die Erstvorstellung. Ich habe es nur reverse engineered. Das Video unten hat 2,5 Minuten zur Generierung benötigt. Also haben wir noch einige Arbeit vor uns. cc: @vllmproject @MiniMaxAI @RyanLeeMiniMax https://github.com/joeynyc/MiniMax-H3-DGX-Spark
mehr auf Arint.info
#AI #DGXSpark #MachineLearning #MiniMaxH3 #OpenSource #TextToVideo #arint_info
-
🎉 Oh joy, someone bought a DGX Spark and decided it's the perfect gadget for playing StarCraft II and "tinkering" with Darktable 😂. Because why not use a $10,000 supercomputer for running games from 20 years ago? 🚀👾
https://daniel.lawrence.lu/blog/2026-07-15-dgx-spark-as-daily-driver/ #DGXSpark #StarCraftII #Tinkering #GamingSupercomputer #TechHumor #HackerNews #ngated -
Come Scegliere il Miglior Mini PC per l'AI Locale nel 2026: Strix Halo vs DGX Spark vs Mac
Un mini PC grande quanto un libro tascabile è oggi in grado di eseguire localmente un modello da 200 miliardi di parametri. Ma la scelta non dipende solo dal prezzo. La capacità della memoria determina quali modelli possono essere caricati, la larghezza di banda influisce sulla velocità di esecuzione e lo stack software — CUDA, ROCm o Metal — stabilisce se gli strumenti che utilizi funzioneranno davvero. Ecco un confronto tra le quattro principali opzioni disponibili nel 2026, con prezzi e benchmark aggiornati.
buysellram.com/blog/how-to-cho…
#AIlocale #LLM #MiniPC #HardwareAI #StrixHalo #DGXSpark #AppleSilicon #EdgeAI #InfrastrutturaAI #Ollama #RyzenAI #AIPC #AMD #NVIDIA #Apple
-
A mini PC the size of a paperback can now run a 200B-parameter model locally. But choosing one isn't about the lowest price tag. Capacity sets what fits, bandwidth sets how fast it runs, and the software stack — CUDA, ROCm, or Metal — decides whether your tools work at all. Here's how the four real options compare in 2026, with current prices and benchmarks.
https://www.buysellram.com/blog/how-to-choose-the-best-mini-pc-for-local-ai-in-2026/
#LocalAI #LLM #MiniPC #AIhardware #StrixHalo #DGXSpark #AppleSilicon #EdgeAI #RyzenAI #AIPC #AMD #NVIDIA #Apple -
Choosing A mini PC isn't about the lowest price tag. Capacity sets what fits, bandwidth sets how fast it runs, and the software stack — CUDA, ROCm, or Metal — decides whether your tools work at all. Here's how the four real options compare in 2026, with current prices and benchmarks.
https://www.buysellram.com/blog/how-to-choose-the-best-mini-pc-for-local-ai-in-2026/
#LocalAI #LLM #MiniPC #AIhardware #StrixHalo #DGXSpark #AppleSilicon #EdgeAI #AIinfrastructure #Ollama #RyzenAI #AIPC #AMD #NVIDIA -
AMD Ryzen AI Max+ 395 vs Nvidia DGX Spark vs Apple Mac — plus when a GPU tower still beats all three. A practical hardware guide for IT managers, developers, and small-business owners weighing a local LLM machine.
Running large language models locally went from a niche hobby to a real procurement question in 2026. A mini PC the size of a paperback can now hold a 200-billion-parameter model — the kind of workload that used to need a server rack.
But picking one isn't about the lowest price. Three things decide whether a model runs well: memory capacity (what fits), memory bandwidth (how fast it runs), and the software ecosystem — CUDA, ROCm, or Metal — that determines whether your existing tools work at all.
There are four real ways to run a local LLM on your desk: a discrete-GPU tower (fastest, but a VRAM wall), AMD Strix Halo mini PCs (big unified memory, cheap, Windows-native), Nvidia's GB10 boxes like the DGX Spark and Dell Pro Max (CUDA, but now $4,699 and Linux-only), and Apple's Mac mini and Mac Studio (high bandwidth, silent, no CUDA).
This guide breaks down which fits which job — with verified specs and current prices.
An appendix at the end collects what early buyers of the AMD “lunchbox” are actually reporting.
https://www.buysellram.com/blog/how-to-choose-the-best-mini-pc-for-local-ai-in-2026/
#LocalAI #LLM #MiniPC #AIhardware #StrixHalo #DGXSpark #AppleSilicon #EdgeAI #AIinfrastructure #Ollama #RyzenAI #AIPC #AMD #NVIDIA #Apple #technology -
A mini PC the size of a paperback can now run a 200B-parameter model locally. But choosing one isn't about the lowest price tag. Capacity sets what fits, bandwidth sets how fast it runs, and the software stack — CUDA, ROCm, or Metal — decides whether your tools work at all. Here's how the four real options compare in 2026, with current prices and benchmarks.
https://www.buysellram.com/blog/how-to-choose-the-best-mini-pc-for-local-ai-in-2026/
#LocalAI #LLM #MiniPC #AIhardware #StrixHalo #DGXSpark #EdgeAI #AIinfrastructure #tech #RyzenAI #NVIDIA #Apple -
NVIDIA DGX Spark 240W電源を徹底分解：デルタ電子が設計、インフィニオンソリューション採用、フルスペックPD3.2の実力 — BigGo ファイナンス https://www.yayafa.com/2821102/ #240W #AgenticAi #AI #ArtificialGeneralIntelligence #ArtificialIntelligence #DGXSpark #GB10GraceBlackwell #MPS #NVIDIA #PD32 #STマイクロエレクトロニクス #インフィニオン #ウェルトトレンド #エージェント型AI #デルタ電子 #人工知能 #汎用人工知能 #炭化ケイ素 #電源アダプター
-
I finally loaded a 120B model - #nemotron3 super, onto my #DGXSpark. With all the stars aligned and goats sacrificed, I think this is the NVFP4 flavour. I'm using it to review patches I made earlier for evaluation.
So far I'm blown away by how _fast_ it is, I'm seeing ~20-25 tokens per second.
It's too soon if this is going to replace my go-to model (qwen3.6-35B-A3B) but I'm looking forward to using during my day job tasks.
I run two models, one on #StrixHalo and one more on the spark. A/B :)