%3 = "tf.Add"(%2, %arg1) : (tensor<?x8x10xf32>, tensor<8x10xf32>) -> tensor<?x8x10xf32>
  %4 = "tf.Add"(%2, %arg2) : (tensor<?x8x10xf32>, tensor<8x10xf32>) -> tensor<?x8x10xf32>
  %5 = "tf.Add"(%arg1, %arg2) : (tensor<8x10xf32>, tensor<8x10xf32>) -> tensor<8x10xf32>

  // CHECK-LABEL: func @parallel_execute_with_tiled_input
  // CHECK-SAME: (%[[ARG_0:[a-z0-9]*]]: tensor<128x10xf32>, %[[ARG_1:[a-z0-9]*]]: tensor<128x10xf32>, %[[ARG_2:[a-z0-9]*]]: tensor<*xi32>, %[[ARG_3:[a-z0-9]*]]: tensor<*xi32>)
  func.func @parallel_execute_with_tiled_input(%arg0: tensor<128x10xf32>, %arg1: tensor<128x10xf32>, %arg2: tensor<*xi32>, %arg3: tensor<*xi32>) -> (tensor<*xi32>, tensor<*xi1>) {
    // CHECK: tf_device.replicate

  %0 = "tf.MatMul"(%arg0, %arg1) {T = "tfdtype$DT_FLOAT", device = "/device:CPU:0", name = "MatMul", transpose_a = true, transpose_b = false} :
(tensor<37x40xf32>, tensor<37x40xf32>) -> tensor<40x40xf32>
  func.return %0 : tensor<40x40xf32>
// CHECK-LABEL: matmul_transposed_a
// CHECK: %[[CST_0:.*]] = arith.constant dense<[1, 0]> : tensor<2xi32>

func.func @foldFill() -> (tensor<3x2x1xf32>, tensor<*xf32>, tensor<*xcomplex<f32>>) {
  %0 = "tf.Const"() {value = dense<[3, 2, 1]> : tensor<3xi32>} : () -> tensor<3xi32>
  %1 = "tf.Const"() {value = dense<23.0> : tensor<f32>} : () -> tensor<f32>
  // CHECK-DAG: "tf.Const"() <{value = dense<2.300000e+01> : tensor<3x2x1xf32>}>
  %2 = "tf.Fill"(%0, %1) : (tensor<3xi32>, tensor<f32>) -> tensor<3x2x1xf32>

  // expected-error @+1 {{'tfl.prelu' op result type '10x10' not broadcast compatible with broadcasted operands's shapes '10x10x10x10'}}
  %0 = "tfl.prelu"(%arg0, %arg1) : (tensor<10x10x10x10xf32>, tensor<10x10x10x10xf32>) -> tensor<10x10xf32>
  func.return %0 : tensor<10x10xf32>
}

// -----