Help using llama.cpp with intel n100?

A user asks for guidance on running llama.cpp server with Gemma 4 E2B on an Intel N100 mini PC running Proxmox.

• The hardware is an Intel N100 mini PC.
• The software stack includes Proxmox and llama.cpp server.
• The model target is Gemma 4 E2B for small tasks.
• The user seeks advice on whether to use CPU only or the iGPU, and which backend to target if using the iGPU.