Function moe_w4a4_grouped_gemm_relu2 

pub fn moe_w4a4_grouped_gemm_relu2(
    gpu: &dyn GpuBackend,
    kernel: KernelHandle,
    a_packed: DevicePtr,
    a_sf: DevicePtr,
    b_packed_ptrs: DevicePtr,
    b_scale_ptrs: DevicePtr,
    scale2_vals: DevicePtr,
    output: DevicePtr,
    expert_offsets: DevicePtr,
    sorted_token_ids: DevicePtr,
    num_experts: u32,
    n_out: u32,
    k: u32,
    max_m_tiles: u32,
    stream: u64,
) -> Result<()>

Grouped W4A4 expert UP GEMM with fused relu^2 (native FP4 tensor cores). A is the pre-quantized NVFP4 latent (packed E2M1 + per-16 E4M3 scales); B comes from the per-expert NVFP4 pointer tables unchanged. Grid: (ceil(n_out/128), max_m_tiles, num_experts) Block: (128, 1, 1)