DGX Spark / GB10 User Forum DGX Spark / GB10
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
|
Introducing NVIDIA NemoClaw
NVIDIA NemoClaw is an open source stack that simplifies running OpenClaw always-on assistants more safely, with a single command. It installs the NVIDIA OpenShell runtime, part of the NVIDIA Agent Toolkit, a secure envir… |
|
0 | 3552 | March 17, 2026 |
|
DGX Spark Performance FAQ
DGX Spark Performance FAQ Performance Blog Check out our blog post on How NVIDIA DGX Spark’s Performance Enables Intensive AI Tasks which details benchmarks using common AI tools. Benchmarking Guide We have released a g… |
|
0 | 3856 | February 2, 2026 |
|
How NVIDIA DGX Spark’s Performance Enables Intensive AI Tasks
Check out our latest tech blog: How NVIDIA DGX Spark’s Performance Enables AI Intensive Tasks |
|
0 | 2438 | November 1, 2025 |
|
DGX Spark Power Clarification
DGX Spark’s peak total system power is 240W. The TDP of the GB10 SOC which includes the GPU and the CPU is 140W. The rest of the system which includes the ConnectX-7, SSD and provisions for USB-C devices is 100W. We have… |
|
0 | 3899 | October 31, 2025 |
|
Welcome! Please Read before Posting
Before posting a new topic, please follow the instructions below Check out our FAQ to see if your questions is answered there: DGX Spark / GB10 FAQ - #7 Search through the other topic and see if someone else has posted … |
|
0 | 1001 | October 28, 2025 |
|
DGX Spark / GB10 FAQ
Initial Setup Q: I am trying to set up my DGX Spark in appliance mode. I am able to connect to the DGX Spark SSID, but it can’t open the webpage it tries to redirect me to (http://spark-xxxx.local). To connect to http://… |
|
0 | 13852 | October 10, 2025 |
|
nvidia/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-NVFP4
|
|
10 | 753 | July 12, 2026 |
|
GLM-5.2 on a 4× GB10 cluster: ~22 tok/s decode, 256K ctx, Recipe
|
|
152 | 7356 | July 12, 2026 |
|
Spark Studio — Open-source inference dashboard for DGX Spark
|
|
15 | 844 | July 12, 2026 |
|
Tokenspeed as alternative to vLLM?
|
|
3 | 1035 | July 12, 2026 |
|
New 2x Spark King? Tencent Hy3 just released
|
|
38 | 2989 | July 12, 2026 |
|
Introducing Spark Auto Round /w OpenCode Instruct dataset
|
|
102 | 3438 | July 11, 2026 |
|
New 2.5x Faster Qwen3.6 NVFP4 Unsloth quants
|
|
12 | 1817 | July 11, 2026 |
|
Is there a Marketplace? I want to get rid of my 1month old Dell GB10 1TB
|
|
13 | 1082 | July 11, 2026 |
|
Question about kvcache size of deepseek4flash
|
|
1 | 47 | July 11, 2026 |
|
Best model for single Spark?
|
|
34 | 3040 | July 11, 2026 |
|
MiMo V2.5 with Dflash FP8 KV 1.5M+ & Deep Speed up to 30 tk/s
|
|
30 | 806 | July 11, 2026 |
|
DeepSeek v4 Flash (Aiden Recipe from Reddit) - 1M token session operational, Cuda 12.1 tailored for DGX Spark GB10
|
|
371 | 17400 | July 11, 2026 |
|
Official NVidia optimized DeepSeek-V4-Flash models?
|
|
28 | 1460 | July 11, 2026 |
|
Easy-vllm — Let a code agent build & serve any model on vLLM for your DGX Sparks
|
|
1 | 97 | July 11, 2026 |
|
Upgrade from ASUS GX10 to Nvidia DGX OS 7.5
|
|
4 | 460 | July 11, 2026 |
|
DGX spark Multi-user
|
|
0 | 83 | July 11, 2026 |
|
4 Node Cluster with CRS504 and 100G Connection: Results
|
|
11 | 415 | July 11, 2026 |
|
Is worth Running NVFP4 Quantization?
|
|
0 | 120 | July 11, 2026 |
|
Dual Spark Ducted Cooling Cage
|
|
278 | 11174 | July 10, 2026 |
|
Nvidia_RTX_Nodes for ComfyUI problem: No nvidia-vfx wheel for DGX Spark
|
|
9 | 630 | July 10, 2026 |
|
Reboot now requires cable out power cycle
|
|
2 | 150 | July 10, 2026 |
|
Root account with no password ond DGX Spark / GB10
|
|
4 | 93 | July 10, 2026 |
|
Two DGX Sparks over the ConnectX-7 direct link — setup notes, throughput numbers,and two questions (federated vs cluster, telemetry)
|
|
2 | 166 | July 10, 2026 |
|
DGX Spark (GB10) GPU clock pinned at 721 MHz under full load — no throttling, not liftable via nvidia-smi
|
|
6 | 181 | July 10, 2026 |