Function moe_expert_gemv 

pub fn moe_expert_gemv(
    gpu: &dyn GpuBackend,
    kernel: KernelHandle,
    input: DevicePtr,
    packed_ptrs: DevicePtr,
    scale_ptrs: DevicePtr,
    scale2_vals: DevicePtr,
    output: DevicePtr,
    expert_indices: DevicePtr,
    n: u32,
    k: u32,
    top_k: u32,
    input_stride: u32,
    stream: u64,
) -> Result<()>

2026-09-25: All top_k expert W4A16 GEMVs in one launch, output [top_k, N]. input_stride is 0 when every expert reads the same input row (gate/up) and K when slot e reads row e (down). A null weight pointer (a remote expert under EP) writes zeros.