Multiple NVIDIA GPUs or Apple Silicon for Large Language Model Inference? - View it on GitHub