For years I’ve been running dietpi in i386 laptops. I’ve decided I will upgrade my setup to run local llms. I am considering to buy the framework desktop which is based in the AMD Ryzen AI Max 395 processor and I am wondering if it will be compatible.
Why I am choosing Dietpi? Simply ease of use. Besides running llms locally I want also to run it as my homelab (torrent client, vpn, plex, pihole, wireguard some docker containers) and in the future some light gaming though streaming (sunshine).
You mean x86_64/amd64? DietPi does not support 32-bit i386 .
It is still x86_64, hence will work. Whether its special AI features (I guess it has an NPU included?) work, is another question. Check whether mainline Linux has a driver for it, and from which version on. Or whether there are at least OSS driver sources that can be compiled via DKMS or so. But the basic hardware features will work.
Graphics/VPU is not the issue, but the NPU probably is. I just mentioned it since you say “AI rig” and refer to the “Ryzen AI” CPU. BitTorrent, VPN/WireGuard, Plex, Pi-hole, Sunshine, all is not an issue, but also all unrelated to AI, not making use of the NPU.
It will be just not available. Usually the NPU driver exposes access e.g.via /dev/dri/render* device node (used by DRM/graphics devices as well) and API, so AI software and libraries can communicate with the NPU. Without the driver AI software cannot use the NPU for hardware-accelerated processing.
However, none of the software you listed makes use of this. It is relevant only if you run local AI/ML models.
.