Comment by irfn

Is there a performance comparison with Ollama. Do both use llama.cpp for serving?