Function forward_dense 

pub fn forward_dense(
    gpu: &dyn GpuBackend,
    k: &Glm5NextMlpKernels,
    cfg: &Glm5NextMlpConfig,
    w: &Glm5NextDenseMlpWeights,
    inter: usize,
    x: DevicePtr,
    out: DevicePtr,
    m: usize,
    ws: &Glm5NextMlpWorkspace,
    stream: u64,
) -> Result<()>

2026-09-25: A BF16 SwiGLU MLP of width inter, down(clamped_swiglu(gate(x), up(x))), for a dense layer or the shared expert. Errors when inter or m does not fit the workspace.