Ollama
Ollama can't split one model across several computers. It spreads a model over the graphics cards inside one machine only; community changes adding llama.cpp RPC have not been merged.
- Mac: No
- NVIDIA PC: No
- AMD PC: No
- Linux, processor only: No
- Raspberry Pi and boards: No
- Windows: No
- Android: No
- iPhone, iPad: No
At a glance
- Best for
- One machine. To cluster, use llama.cpp RPC (same GGUF files) or another tool here.
- Runs on
- No multi-machine support on any device. LM Studio's LM Link and NVIDIA's Personal AI Router are similar: they let you use a model on another of your machines, but each model still runs on one machine.
- How it splits a model
- None across machines.
Limits to know
- No clustering.
Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.