Function emit_qwen30b_generate_batched 

pub fn emit_qwen30b_generate_batched(
    cfg: &Qwen30bCfg,
    wdir: &str,
    prompt: &[usize],
    n_gen: usize,
    batch: usize,
) -> String

Emit a self-contained CUDA program that runs the real Qwen3-30B-A3B decode over batch parallel sequences and reports aggregate S_TG plus the golden token stream for every sequence. Weights fopen from wdir.