Conversation

Notices

1. Embed this notice
   Mauricio Teixeira 🇧🇷🇺🇲 (badnetmask@hachyderm.io)'s status on Friday, 28-Feb-2025 10:02:50 JST Mauricio Teixeira 🇧🇷🇺🇲

   I spent a few days thinking whether I should upgrade an SFF PC I use as a server, or buy a mini/tiny. The problem in the end boiled down to a decision between a tiny with an AMD APU, or adding an NVIDIA to the SFF. Sadly support for APU/NPU on Linux is still very unclear/uncertain right now, and I lack the time and skills to contribute to the development. So, let's go with a used NVIDIA for now.

   (For now... I'm sure I'm gonna regret it soon.)

   #HomeLab #SelfHosted

   • Embed this notice
     Mauricio Teixeira 🇧🇷🇺🇲 (badnetmask@hachyderm.io)'s status on Saturday, 08-Mar-2025 13:06:21 JST Mauricio Teixeira 🇧🇷🇺🇲

     in reply to

     As expected, I regretted my attempt to upgrade my SFF server. None of the parts worked, so as long as the vendors would accept them back, all I wasted was time.

     What's next: decide between a tower PC with a regular GPU, or a modern mini PC with an APU/NPU.

     I'm actually open to suggestions of mini PCs that can run Ollama.

     #HomeLab #SelfHosted #HomeAssistant #Ollama

   • Embed this notice
     Mauricio Teixeira 🇧🇷🇺🇲 (badnetmask@hachyderm.io)'s status on Saturday, 08-Mar-2025 21:54:14 JST Mauricio Teixeira 🇧🇷🇺🇲

     in reply to

     • Santiago

     @sanzky
     No Mac for me. But thanks.

   • Embed this notice
     Santiago (sanzky@masto.es)'s status on Saturday, 08-Mar-2025 21:54:15 JST Santiago

     in reply to

     @badnetmask a mac mini is the smallest thing I that I know that can run ollama