Benchmarking Gemma 4 E2B on CPU with llama.cpp: A Practical Local AI Experiment

1. Why This Experiment? I want to run a local language model for everyday office work...