// CHECK-LABEL:   func @batchmatmulv2_lhs_batch
// CHECK:           "mhlo.dynamic_broadcast_in_dim"({{.*}}, {{.*}}) <{broadcast_dimensions = dense<[1, 2]> : tensor<2xi64>}>
// CHECK:           "mhlo.dot_general"({{.*}}, {{.*}}) <{
// CHECK-SAME:        lhs_batching_dimensions = [0]
// CHECK-SAME:        rhs_batching_dimensions = [0]
// CHECK-SAME:        lhs_contracting_dimensions = [2]

	a1, _ := strconv.ParseUint(fmt.Sprintf("%p", &b1[1])[2:], 16, 64)
	if a1 != a0+1 {
		fmt.Println("FAIL")
		if report {
			fmt.Println("alignment should be 1, is", a1-a0)
		}
		status = 1
	}
	var b2 [10]T2
	a0, _ = strconv.ParseUint(fmt.Sprintf("%p", &b2[0])[2:], 16, 64)
	a1, _ = strconv.ParseUint(fmt.Sprintf("%p", &b2[1])[2:], 16, 64)
	if a1 != a0+2 {
		if status == 0 {
			fmt.Println("FAIL")

tensor<f64>, tensor<!tf_type.resource>)

# CHECK-SUBTYPE: func @main(%arg0: tensor<10xi32>, %arg1: tensor<!tf_type.variant<tensor<10xf32>>>, %arg2: tensor<10xi8>, %arg3: tensor<10xf64>, %arg4: tensor<!tf_type.resource<tensor<10xi32>>>) -> (tensor<10xi32>, tensor<!tf_type.variant<tensor<10xf32>>>, tensor<10xi8>, tensor<10xf64>, tensor<!tf_type.resource<tensor<10xi32>>>)...

  // CHECK-LABEL: func @init_all_tables_with_explicit_vocab_size
  // CHECK-DAG: [[CST:%.*]]  = arith.constant dense<["apple", "banana"]> : tensor<2x!tf_type.string>
  // CHECK-DAG: [[CST_0:%.*]]  = arith.constant dense<[0, 1]> : tensor<2xi64>
  // CHECK: [[VAL:%.*]] = "tf.HashTableV2"()
  // CHECK: "tf.LookupTableImportV2"([[VAL]], [[CST]], [[CST_0]])

  // CHECK-LABEL: func @init_all_tables_with_explicit_vocab_size
  // CHECK-DAG: [[CST:%.*]]  = arith.constant dense<["apple", "banana"]> : tensor<2x!tf_type.string>
  // CHECK-DAG: [[CST_0:%.*]]  = arith.constant dense<[0, 1]> : tensor<2xi64>
  // CHECK: [[VAL:%.*]] = "tf.HashTableV2"()
  // CHECK: "tf.LookupTableImportV2"([[VAL]], [[CST]], [[CST_0]])

    func.func @main() -> () {
      %cst0 = "tf.Const"(){ value = dense<0> : tensor<3x5xi1>} : () -> tensor<3x5xi1>
      %0 = "tf.Where"(%cst0) : (tensor<3x5xi1>) -> tensor<?x2xi64>
      func.return
    }
  })";

  CellReader<int64_t> dynamism_counter(kDynamismOpCounterStreamzName);
  CellReader<int64_t> dynamism_function_counter(
      kDynamismFunctionCounterStreamzName);

  %7 = "mhlo.broadcast_in_dim"(%3) <{broadcast_dimensions = dense<3> : tensor<1xi64>}> : (tensor<5xf32>) -> tensor<32x28x28x5xf32>
  %8 = mhlo.add %6, %7 : tensor<32x28x28x5xf32>
  %9 = mhlo.maximum %8, %1 : tensor<32x28x28x5xf32>
  %10 = "mhlo.reshape"(%9) : (tensor<32x28x28x5xf32>) -> tensor<32x3920xf32>

}

// Decompose tf.RngReadAndSkip.
//
// For Philox, the resource variable holds a tensor<3xi64> with the state:
//   [counter_lo, counter_hi, key]
//
//   RngReadAndSkip increments the 128 bit counter value by 256 * delta and
//   returns the original state value.
//
// For Threefry, the resource variable holds a tensor<2xi64> with the state:
//   [counter, key]
//