}

// CHECK-LABEL: add
func.func @add(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) -> tensor<1xf32> {
  %0 = "tf.AddV2"(%arg0, %arg1) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
  func.return %0: tensor<1xf32>
// CHECK: %[[ADD_0:.*]] = "tf.AddV2"(%arg0, %arg1) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
// CHECK: return %[[ADD_0]] : tensor<1xf32>
}

// CHECK-LABEL: softmax

    %5 = "quantfork.qcast"(%4) {volatile} : (tensor<1x3xf32>) -> tensor<1x3x!quant.uniform<i8:f32, 0.13170163023705575:-1>>
    %6 = "quantfork.dcast"(%5) : (tensor<1x3x!quant.uniform<i8:f32, 0.13170163023705575:-1>>) -> tensor<1x3xf32>
    %7 = stablehlo.reshape %6 : (tensor<1x3xf32>) -> tensor<3x1xf32>

  // CHECK: device = "cpu"
  %2 = "tf.Relu"(%1) {T = f32, _output_shapes = ["tfshape$dim { size: 3 } dim { size: 3 }"], device = "cpu"} : (tensor<3x3xf32>) -> tensor<3x3xf32>
  // CHECK: device = "xpu"
  %3 = "tf.Relu"(%2) {T = f32, _output_shapes = ["tfshape$dim { size: 3 } dim { size: 3 }"]} : (tensor<3x3xf32>) -> tensor<3x3xf32>
  func.return %3 : tensor<3x3xf32>
}

  func.return %2 : tensor<3x3xf32>

func.func @pack(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) -> tensor<2x1xf32> {
  %0 = "tfl.pack"(%arg0, %arg1) {axis = 0 : i32, values_count = 2 : i32} : (tensor<1xf32>, tensor<1xf32>) -> tensor<2x1xf32>
  func.return %0 : tensor<2x1xf32>
}

// CHECK:   func @pack(%[[VAL_0:.*]]: tensor<1xf32>, %[[VAL_1:.*]]: tensor<1xf32>) -> tensor<2x1xf32> {

# CHECK-NEXT: [[R0:%.*]] = "tf.ReadVariableOp"([[ARG1]]) {{{.*}}} : (tensor<!tf_type.resource<tensor<1x3xf32>>>) -> tensor<1x3xf32>
# CHECK-NEXT: [[R1:%.*]] = "tf.MatMul"([[ARG0]], [[R0]]) <{{{.*}}}> {device = ""} : (tensor<3x1xf32>, tensor<1x3xf32>) -> tensor<3x3xf32>
# CHECK-NEXT: return [[R1]] : tensor<3x3xf32>


def Test():

  x = tf.constant([[1.0], [1.0], [1.0]])
  y = tf.compat.v1.get_variable(
      name='y',

func.func @device_test(%arg0: tensor<3x1xf32>) -> (tensor<3x3xf32>) {

  // CHECK: device = "gpu"
  %0 = "tf.Const"() {value = dense<[[1.0, 2.0, 3.0]]> : tensor<1x3xf32>} : () -> tensor<1x3xf32>
  // CHECK: device = "gpu"
  %1 = "tf.MatMul"(%arg0, %0) {T = f32, _output_shapes = ["tfshape$dim { size: 3 } dim { size: 3 }"], device = "", transpose_a = false, transpose_b = false} : (tensor<3x1xf32>, tensor<1x3xf32>) -> tensor<3x3xf32>
  // CHECK: device = "cpu"

  // CHECK: %[[ADD:.*]] = tfl.add %[[VAR_VAL]], %arg0 {fused_activation_function = "NONE"} : tensor<1x10xf32>
  // CHECK: "tfl.assign_variable"(%[[RESOURCE]], %[[ADD]]) : (tensor<!tf_type.resource<tensor<1x10xf32>>>, tensor<1x10xf32>) -> ()
  // CHECK: %[[RESULT:.*]] = "tfl.read_variable"(%[[RESOURCE]]) : (tensor<!tf_type.resource<tensor<1x10xf32>>>) -> tensor<1x10xf32>

module {
  func.func @simpleTest(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>, %arg2: tensor<1xf32>, %arg3: tensor<1xf32>) -> tensor<2x1xf32> {
    %0 = func.call @func_0_GPU_FLOAT(%arg0, %arg1, %arg2) {tac.interface_name = "func_0"} : (tensor<1xf32>, tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    %1 = func.call @func_1_GPU_FLOAT(%arg0, %arg3) {tac.interface_name = "func_1"} : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>

  %1 = "tf.BroadcastTo"(%arg1, %cst) : (tensor<5x1xf32>, tensor<2xi64>) -> tensor<5x7xf32>
  %2 = "tf.Add"(%0, %1) : (tensor<5x7xf32>, tensor<5x7xf32>) -> tensor<5x7xf32>
  func.return %2 : tensor<5x7xf32>
  // CHECK: %[[V0:.*]] = "tf.Add"(%arg0, %arg1) : (tensor<7xf32>, tensor<5x1xf32>) -> tensor<5x7xf32>
  // CHECK: %[[V0]] : tensor<5x7xf32>
}

// CHECK-LABEL: @broadcast_batch_matmul_v2_rhs