I’m looking to build a low-end ollama LLM server to improve home assistant voice control, Immich image recognition and a few other services. With the current cost of hardware components like memory, I’m looking to build something small, but somewhat expandable.
I have an old micro-atx form factor computer that I’m thinking will be a good option to upgrade. I’d love recommendations on motherboards, processors, and video card combos that would likely be compatible and sufficient to run a decent server while keeping costs lower, basically, the best bang for the buck. I have a couple of M.2 SSDs I can re-purpose. Would prefer the motherboard has 2.5Gbit Ethernet, but otherwise I’m open.
Also recommendations on sites to purchase good quality memory at reasonable prices that ship to the US. I’d be willing to look at lightly used components, too.
Any advice on any of these topics would be greatly appreciated. The advice I’ve found has all been out of date especially with crypto fading so video cards are not as expensive, but LLM data centers eating up and reserving memory before it’s even manufactured.
So just glancing at the site, are these basically laptop CPU and RAM parts just packaged in a desktop form-factor case and that’s why they’re soldered? Seems like they also don’t have much expansion capability much like a laptop such as only having a single PCI-E x4 slot with a proprietary connection interface, so I couldn’t later add a graphics card for example. Unless, I’m just missing something, and if so please let me know.
Either way thanks for letting me know about the option.
The main benefit is the strix halo cpu uses unified memory, thats why it’s soldered, not bc it uses laptop parts
Ok, so short, wide bus from CPU to memory? Makes sense. I didn’t really mean the CPU so much as the main board is very laptop like. Very little expansion capabilities other than external connectors like audio, Ethernet, etc., but no ability to add functional or incremental upgrades like a GPU or an additional stick of memory respectively.
The point of the stays halo series is the unified memory, so an additional GPU wouldn’t be very useful, no?
GPU has longer more specialized processing that is significantly better at doing certain things than a generalized CPU. Things that benefits from tons of parallel threads more than being able to react more quickly to changing conditions will process significantly faster on a GPU than a CPU. Having the ability to add more parallel processing units (I.e. additonal GPUs is a huge advantage in crypto, machine learning analysis, video and 3D polygon rendering, etc. Basically things that have tons of work to get done, and that set of work is unlikely to be affected by the outcome of work running in parallel. A CPU needs ro balance that fact that it wastes a ton of processing if it tries to calculate something that is more dynamic. The content of the pixels in a video, or needing to process the entire set of millions of relationships to reach a conclusion will benefit from the GPU and adding more GPUs in parallel benefits more because most of those workloads will always be extremely large multiples of the number of threads in a single GPU. In a CPU, things like unpredictability of human input or branching logic that has too many possible branches to just calculate them all in the hopes that one of those branches might be the one that actually gets used, so the threads calculate much smaller units and adding more CPU cores/threads then has greatly decreasing return on investment because either you have to wait for something to actually happen or you haven’t calculate all the most probable outcome and waste a ton of work. The shorter threads mean less efficient at complex calculation, but getting results that some other process is waiting in faster.
Anyway, there’s a reason we still have GPUs and dont just keep increasing CPU cores instead.