Phi-4 on 18GB M3 Macbook Pro | CPU vs GPU

Описание: Comparing GPU vs CPU Performance on MacBook Pro M3 - Microsoft's Phi-4 Language Model Test

In this technical demo, I test the performance differences between GPU and CPU processing using Microsoft's Phi-4 small language model (9B parameters) running on Ollama both within Docker and directly installed on the laptop. The video shows real-time GPU and CPU monitoring while running identical text generation tasks on both processors. Results demonstrate significantly faster performance when utilizing the MacBook Pro M3's GPU compared to CPU processing, with metrics and timestamps shown.

For GPU setup instructions, check out my previous tutorial on configuring N1 for GPU usage:    • Quick Guide to Ollama & N8N on Apple GPU