Function moe_w4a16_grouped_gemm_ptrtable_n128 

pub fn moe_w4a16_grouped_gemm_ptrtable_n128(
    gpu: &dyn GpuBackend,
    kernel: KernelHandle,
    a: DevicePtr,
    b_packed_ptrs: DevicePtr,
    b_scale_ptrs: DevicePtr,
    scale2_vals: DevicePtr,
    c: DevicePtr,
    expert_offsets: DevicePtr,
    sorted_token_ids: DevicePtr,
    num_experts: u32,
    n_out: u32,
    k: u32,
    max_m_tiles: u32,
    stream: u64,
) -> Result<()>

Pointer-table grouped GEMM with N_TILE=128 (transposed or wide kernels).

Grid: (ceil(n_out/128), max_m_tiles, num_experts) Block: (128, 1, 1)