System Setup Guide

🧰 Hardware Specifications

Mini PC: UM890 PRO Minisforum
RAM: 96GB Crucial DDR5 (2x48GB) 5600MHz SODIMM - CL46 - CT2K48G56C46S5
Storage: M.2 NVMe Samsung SSD 980 PRO 1TB (Model: 5B2QGXA7)
GPU: PNY GeForce RTX™ 5060 Ti 16GB ARGB Overclocked Triple Fan DLSS 4
eGPU Docking: GTBOX G-DOCK with OCuLink USB4 and Integrated 800W Huntkey Power Supply

📌 Installation Tips

OS: Install Ubuntu 24.04 LTS desktop edition
HDMI: Use the Mini PC's integrated HDMI (do not connect NVIDIA GPU HDMI output)
Drivers: Enable Ubuntu non-free repository during install for all drivers

Post-Install: Run these commands:

sudo apt update && sudo apt upgrade -y
sudo reboot

🧪 NVIDIA Driver Installation

🔗 Official Instructions

Install CUDA Toolkit 13.0 for Ubuntu 24.04

🛠️ Command Steps (at the time of writing, the instructions are)

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install cuda-toolkit-13-0

🚀 Install drivers

sudo apt-get install -y nvidia-open

🧠 Open vs Proprietary Drivers

For NVIDIA Grace Hopper or Blackwell platforms, only open-source drivers are supported:

"For cutting-edge platforms such as NVIDIA Grace Hopper or NVIDIA Blackwell, you must use the open-source GPU kernel modules. The proprietary drivers are unsupported on these platforms."
Source

🧠 Ollama Installation

curl -fsSL https://ollama.com/install.sh | sh

📦 Output

>>> Installing ollama to /usr/local
>>> Downloading Linux amd64 bundle
######################################################################## 100.0%
>>> Creating ollama user...
>>> Adding ollama user to render group...
>>> Adding ollama user to video group...
>>> Adding current user to ollama group...
>>> Creating ollama systemd service...
>>> Enabling and starting ollama service...
Created symlink /etc/systemd/system/default.target.wants/ollama.service → /etc/systemd/system/ollama.service.
>>> NVIDIA GPU installed.

🐳 Docker & Portainer Setup

📌 Next Steps

Power off the system
Connect eGPU using OCuLink cable, power on dock station and mini PC
If you see the login screen:
- Log in and run:
```
nvidia-smi
```
  (If output appears, everything is working)
Run Ollama and open WebUI in Docker

🚀 Run Open WebUI

docker run -d --restart always --network=host \
  -e OLLAMA_BASE_URL=http://localhost:11434 \
  -v open-webui:/app/backend/data \
  --name open-webui \
  ghcr.io/open-webui/open-webui

🌐 Access Open WebUI

Open your browser and go to:

http://localhost:8080

✅ Final Step

Install at least one model:

ollama pull mistral-small3.2

Now you can use Open WebUI and enjoy! 🎉

🔍 Notes

Ensure the eGPU is properly connected before powering on
Verify nvidia-smi output confirms GPU detection
Use OLLAMA_BASE_URL for Ollama service compatibility

ZEROF/Setting AI Home Lab.md