// 1. InferSpmd (Infer DistAttr of Inputs&Outputs)
    auto meta_dist_x = MakeDistMetaTensor(*x.impl());
    auto meta_dist_y = MakeDistMetaTensor(*y.impl());
    auto spmd_info = phi::distributed::MatmulSpmdInferForward(meta_dist_x, meta_dist_y, transpose_x, transpose_y);

  // Auto Parallel condition
  if (AllInputsAreDistTensor(x, y)) {
    // 1. InferSpmd (Infer DistAttr of Inputs&Outputs)
    auto meta_dist_x = MakeDistMetaTensor(*x.impl());
    auto meta_dist_y = MakeDistMetaTensor(*y.impl());
    auto spmd_info = phi::distributed::MatmulSpmdInferForward(meta_dist_x, meta_dist_y, transpose_x, transpose_y);
    // 2. Create API Output & Prepare Dist and Dense Output
    Tensor api_output;
    auto dist_out = SetKernelDistOutput(&api_output, spmd_info.second[0]);
    auto dense_out = dist_out->unsafe_mutable_value();
    // 3. Infer DistTensor's Global Shape
    phi::MetaTensor meta_dist_out(dist_out);
    phi::MatmulInferMeta(meta_dist_x, meta_dist_y, transpose_x, transpose_y, &meta_dist_out);
    // 4. Select Kernel
    VLOG(6) << "matmul API dist branch: kernel key: [" << kernel_backend << ", " << kernel_layout << ", "<< kernel_data_type << "]";
    auto kernel_result = phi::KernelFactory::Instance().SelectKernelOrThrowError(
        "matmul", {kernel_backend, kernel_layout, kernel_data_type});
    const auto& kernel = kernel_result.kernel;
    VLOG(6) << "matmul kernel: " << kernel;
    auto* dev_ctx = GetDeviceContextByBackend(kernel_result.has_fallback_cpu ? Backend::CPU : kernel_backend);
    // 5. Reshard Input
    auto dist_input_x = ReshardDistTensor(dev_ctx, x, spmd_info.first[0]);
    auto dist_input_y = ReshardDistTensor(dev_ctx, y, spmd_info.first[1]);
    // 6. PrepareData (DataTransform & Prepare Dense Input)
    dist_input_x = PrepareDataForDistTensor(dist_input_x, GetKernelInputArgDef(kernel.InputAt(0), kernel_backend), {}, kernel_result.is_stride_kernel);
    auto input_x = &dist_input_x->value();
    dist_input_y = PrepareDataForDistTensor(dist_input_y, GetKernelInputArgDef(kernel.InputAt(1), kernel_backend), {}, kernel_result.is_stride_kernel);
    auto input_y = &dist_input_y->value();
    // 7. Infer Local DenseTensor Meta
    phi::MetaTensor meta_dense_out(dense_out);
    phi::MatmulInferMeta(MakeMetaTensor(*input_x), MakeMetaTensor(*input_y), transpose_x, transpose_y, &meta_dense_out);
    // 8. DenseTensor Kernel Call
    using kernel_signature = void(*)(const phi::DeviceContext&, const phi::DenseTensor&, const phi::DenseTensor&, bool, bool, phi::DenseTensor*);
    auto* kernel_fn = kernel.GetVariadicKernelFn<kernel_signature>();
    (*kernel_fn)(*dev_ctx, *input_x, *input_y, transpose_x, transpose_y, dense_out);
    // 9. Return
    return api_output;
  }