Ollama CUDA detection and GPU verification
Parent: Thunderbolt eGPU on Linux for local LLM inference · Topic entry
A published reference is not available for this topic yet.
Also known as: ollama, inference compute, library=CUDA, ollama ps, 100% GPU, llama-server, tokens/s, OLLAMA_IGPU_ENABLE, warm-up
Children
- No children recorded.