Building a local AI server
ai
I have been a avid tinkerer with home pc’s. I enjoy building pc. Now that LLM’s are a necessity for a developer, I wanted to build a great performant system for myself.
Requirements
Hardware
My requirements were quite basic:
- A minimum of 16GB VRAM (preferably Nvidia)
- A minimum of 16GB RAM
- A minimum of 6 cores / 12 threads Ryzen CPU
- A motherboard with 2 PCI GPU slots (does not need to be 16 lanes PCI)
- A minimum of 600 watt power supply
- A wi-fi smart-plug with scheduling capabilities
- A motherboard which support power-on after power restoration
Software
- Linux OS with a long-term support and used by a lot of users
- Linux OS with Tailscale native support (systemd)
- Linux OS with easy installation of Docker, Nvidia Container Toolkit and Cuda Toolkit
- Linux OS with OpenSSH server to manage server remotely
- Linux OS with crontab to schedule running of scripts
Server
I ended up purchasing a second-hand PC with all the above requirements and then purchased two RTX 3060 GPUs with 12 GB Vram each. The RAM capacity is somewhat low for this use-case and I plan to upgrade it in the future.
Scheduling
Updates
As everything runs on an Ubuntu LTS 24.04.