ci : add test-llama-archs tensor split for Metal (#27598)

Run test-llama-archs with 1 to 4 GGML_METAL_DEVICES, mirroring the
existing CUDA runs, and dispatch the job unconditionally since the
per-backend guards now decide what to run.

Assisted-by: llama.cpp:DeepSeek-v4-Flash-0731