// CHECK-LABEL: broadcast_to_to_reshape_i64
func.func @broadcast_to_to_reshape_i64(%arg0: tensor<4x4x4xf32>, %arg1 : tensor<4xi64>) -> tensor<1x4x4x4xf32> {
  %0 = "tfl.broadcast_to"(%arg0, %arg1) : (tensor<4x4x4xf32>, tensor<4xi64>) -> tensor<1x4x4x4xf32>
  // CHECK: "tfl.cast"
  // CHECK-SAME: (tensor<4xi64>) -> tensor<4xi32>
  // CHECK-NEXT: "tfl.reshape"
  // CHECK-SAME: (tensor<4x4x4xf32>, tensor<4xi32>) -> tensor<1x4x4x4xf32>

// test stablehlo roundtrip

// Identity function to make the exporter happy
func.func @main(%arg0: tensor<4xi8>) -> tensor<4xi8> {
  func.return %arg0 : tensor<4xi8>
}

//CHECK:func.func @main(%arg0: tensor<4xi8>) -> tensor<4xi8> attributes {tf.entry_function = {inputs = "arg0", outputs = "arg0"}} {
//CHECK: return %arg0 : tensor<4xi8>
//CHECK:}

func.func @logistic(%arg0: tensor<1x1x1x96xf32>) -> tensor<1x1x1x96xf32> {

  %1 = mhlo.constant dense<[[[[0, 1, 2, 3]], [[0, 1, 2, 3]]]]> : tensor<1x2x1x4xi32>
  %2 = "mhlo.broadcast_in_dim"(%0) <{broadcast_dimensions = dense<[0, 1, 2, 3]> : tensor<4xi64>}> : (tensor<1x1x1x4xi32>) -> tensor<1x2x1x4xi32>
  %3 = mhlo.multiply %1, %2 : tensor<1x2x1x4xi32>
  // CHECK:      return %[[RES]] : tensor<1x2x1x4xi32>
  return %3 : tensor<1x2x1x4xi32>
}

		uint64(s[9]))<<4|
		uint64(s[10]))<<4|
		uint64(s[11]))<<4|
		uint64(s[12]))<<4|
		uint64(s[13]))<<4|
		uint64(s[14]))<<4 |
		uint64(s[15]))
	return r
}

// Compute the determinant of a 4x4-matrix by the sum
// over all index permutations.
func determinant(m [4][4]float64) float64 {
	return m[0][0]*m[1][1]*m[2][2]*m[3][3] -
		m[0][0]*m[1][1]*m[2][3]*m[3][2] -
		m[0][0]*m[1][2]*m[2][1]*m[3][3] +

// operate on quantized values.
//
// Examples from storage to quantized type:
//   i8 -> !quant<"uniform[i8:f32]{1.0}">
//   tensor<4xi8> -> tensor<4x!quant<"uniform[i8:f32]{1.0}">>
//   vector<4xi8> -> vector<4x!quant<"uniform[i8:f32]{1.0}">>
def Quantization_StorageCastOp : Quantization_Op<"scast", [Pure]> {
  let arguments = (ins quant_RealOrStorageValueType:$arg);

  %values = "tf.Const"() {value = dense<[1, 2, 3, 4]> : tensor<4xi64>} : () -> tensor<4xi64>
  "tf.LookupTableImportV2"(%handle, %keys, %values) {device = ""} : (tensor<!tf_type.resource>, tensor<4x!tf_type.string>, tensor<4xi64>) -> ()

  func.return %2 : tensor<?x?x?x?xf32>

// CHECK-DAG: %[[CST:.*]] = arith.constant dense<0> : tensor<4xi64>
// CHECK-DAG: %[[CST0:.*]] = arith.constant dense<[1, 9, 23, 768]> : tensor<4xi64>
// CHECK-DAG: %[[CST1:.*]] = arith.constant dense<1> : tensor<4xi64>

// operate on quantized values.
//
// Examples from storage to quantized type:
//   i8 -> !quant<"uniform[i8:f32]{1.0}">
//   tensor<4xi8> -> tensor<4x!quant<"uniform[i8:f32]{1.0}">>
//   vector<4xi8> -> vector<4x!quant<"uniform[i8:f32]{1.0}">>
def quantfork_StorageCastOp : quantfork_Op<"scast", [Pure]> {
  let arguments = (ins quant_RealOrStorageValueType:$arg);

func.return %0 : tensor<2x?x2x4xf32>

// CHECK-LABEL:     convert_dot_general_dynamic_batch_dim
// CHECK:           %0 = "tfl.pseudo_const"() <{value = dense<[0, 1, 3, 2]> : tensor<4xi64>}> : () -> tensor<4xi64>
// CHECK-NEXT:      %1 = "tfl.cast"(%0) : (tensor<4xi64>) -> tensor<4xi32>
// CHECK-NEXT:      %2 = "tfl.transpose"(%arg1, %1) : (tensor<2x?x4x3xf32>, tensor<4xi32>) -> tensor<2x?x3x4xf32>

// See https://cr.yp.to/chacha/chacha-20080128.pdf.
//
// ChaCha8 operates on a 4x4 matrix of uint32 values, initially set to:
//
//	const1 const2 const3 const4
//	seed   seed   seed   seed
//	seed   seed   seed   seed
//	counter64     0      0
//
// We use the same constants as ChaCha20 does, a random seed,
// and a counter. Running ChaCha8 on this input produces
// a 4x4 matrix of pseudo-random values with as much entropy
// as the seed.
//