• 0 Posts
  • 22 Comments
Joined 3 years ago
cake
Cake day: July 4th, 2023

help-circle





  • While I get leaning towards AMD products, I’ve been doing so as well, when I built my first server with a Ryzen 5 2400GE I have found that there just isn’t as much resources/support for enabling transcoding with the vega 11 in Jellyfin or Immich. Most Intel iGPU’s have a hardware chip specifically tuned for transcoding called quicksync that you should strongly consider.

    Especially in the $100-200 price range tiny mini micro’s from HP/Lenovo/Dell are widely available and offer lots of capability in a power-efficient (~10-15w idle, 40-50w full load) and easily maintainable form factor. The Lenovo’s in particular are interesting due to a few models having full pci-e slots if you decide later you want a GPU.
    Lenovo pci-e

    Finally for software I would suggest looking into Cosmos Cloud, I use it and have found it made it so much easier to setup and manage all my docker containers and domain name/reverse proxy settings.






  • I’m actually right there with you, I have a 3060 12gb and tbh I think it’s the absolute most cost effective GPU option for home use right now. You can run 14B models at a very reasonable pace.
    Doubling or tripling the cost and power draw just to get 16-24gb doesn’t seem worth it to me. If you really want an AI-optimized box I think something with the new Ryzen Max chips would be the way to go - like an ASUS ROG Z-Flow, Framework Desktop or the GMKtek option whatever it’s called. Apple’s new Mac Minis are also great options. Both Ryzen Max and Apple make use of shared CPU/GPU memory so you can go up 96GB+ at much much lower power draws.







  • My first time I started winging it with a raspberry pi, docker, and nginx and it took me like 2 months to get one service up and running and I didn’t feel it was very secure - fail2ban didn’t work, geoblocking didn’t work, and updates were manual.

    When I re-started from scratch with an x86 device and cosmos it has been shockingly easy in comparison. Not only is it much quicker to spool the service up (app store), they can be automatically updated, the proxy has options for geoblocking, rate limiting, etc.

    I’ve even got some of the services below built from a custom compose file instead of the app store, some use remote storage and some are set up with OAuth SSO. There’s still mild troubleshooting for a lot of things but it’s been much easier for me to understand and fix issues, plus there’s an active discord community as well.