%0 = "tf.Mul"(%arg1, %arg0) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    func.return %0 : tensor<1xf32>
  }
// CHECK: func private @mul1(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) -> tensor<1xf32> attributes {tf.entry_function = {inputs = "mul1_y:0,mul1_x:0", outputs = "PartitionedCall:0"}}
// CHECK:   %[[MUL_0:.*]] = "tf.Mul"(%arg1, %arg0) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>

loop:
	CBZ	R0, done

	LDP.P	16(R2), (R5, R6)
	LDP.P	16(R2), (R7, R8)

	LDP	(R1), (R9, R10)
	ADDS	R4, R9
	MUL	R6, R3, R14
	ADCS	R14, R10
	MUL	R7, R3, R15
	LDP	16(R1), (R11, R12)
	ADCS	R15, R11
	MUL	R8, R3, R16
	ADCS	R16, R12
	UMULH	R8, R3, R20
	ADC	$0, R20

	MUL	R5, R3, R13
	ADDS	R13, R9
	UMULH	R5, R3, R17
	ADCS	R17, R10
	UMULH	R6, R3, R21
	STP.P	(R9, R10), 16(R1)
	ADCS	R21, R11

            tensor_shape {
            }
            int_val: 2
          }
        }
      }
    }
    node_def {
      name: "mul"
      op: "Mul"
      input: "args_0"
      input: "mul/y:output:0"
      attr {
        key: "T"
        value {
          type: DT_INT32
        }
      }
    }
    node_def {
      name: "Identity"
      op: "Identity"

			u := get(U)
			if end(u) != 0 {
				done = true
			}
			Z.dat <- mul(i2tor(1, int64(i)), u)
		}
		Z.dat <- finis
	}(c, U, Z)
	return Z
}

// Binomial theorem (1+x)^c

func Binom(c *rat) PS {
	Z := mkPS()
	go func(c *rat, Z PS) {
		n := 1
		t := itor(1)
		for c.num != 0 {
			put(t, Z)
			t = mul(mul(t, c), i2tor(1, int64(n)))
			c = sub(c, one)
			n++
		}

  func.return %2 : tensor<4xf32>

// CHECK-NEXT: %[[CST:.*]] = arith.constant dense<1.000000e+00> : tensor<4xf32>
// CHECK-NEXT: %[[MUL:.*]] = tfl.mul %arg0, %[[CST]] {fused_activation_function = "NONE"} : tensor<4xf32>
// CHECK-NEXT: %[[CUSTOM_1:.*]] = "tfl.custom_tf"(%[[MUL]], %[[CST]]) ({
// CHECK-NEXT: ^bb0(%arg1: tensor<4xf32>, %arg2: tensor<4xf32>):

	t1.Mul(t4, y3)                            // t1 := t4 * Y3
	t2.Mul(t0, y3)                            // t2 := t0 * Y3
	y3.Mul(x3, z3)                            // Y3 := X3 * Z3
	y3.Add(y3, t2)                            // Y3 := Y3 + t2
	x3.Mul(t3, x3)                            // X3 := t3 * X3
	x3.Sub(x3, t1)                            // X3 := X3 - t1
	z3.Mul(t4, z3)                            // Z3 := t4 * Z3

			x4 := new(Rat).SetInt(new(Int).Binomial(int64(i+j), int64(i)))

			x1.Mul(x1, x2)
			x1.Mul(x1, x3)
			x1.Mul(x1, x4)
			x1.Mul(x1, x4)

			if (i+j)&1 != 0 {
				x1.Neg(x1)
			}

			a.set(i, j, x1)
		}
	}
	return a
}

func (a *matrix) mul(b *matrix) *matrix {
	if a.m != b.n {
		panic("illegal matrix multiply")
	}
	c := newMatrix(a.n, b.m)

}

func TestMulUintptr(t *testing.T) {
	for _, test := range mulUintptrTests {
		a, b := test.a, test.b
		for i := 0; i < 2; i++ {
			mul, overflow := MulUintptr(a, b)
			if mul != a*b || overflow != test.overflow {
				t.Errorf("MulUintptr(%v, %v) = %v, %v want %v, %v",
					a, b, mul, overflow, a*b, test.overflow)
			}
			a, b = b, a
		}
	}
}

var SinkUintptr uintptr
var SinkBool bool

var x, y uintptr

// CHECK-LABEL: broadcast_to_bf16
// CHECK:  [[CST:%.*]] = arith.constant dense<1.000000e+00> : tensor<3x3xbf16>
// CHECK:  [[MUL:%.*]] = tfl.mul(%arg0, [[CST]]) <{fused_activation_function = "NONE"}> : (tensor<3xbf16>, tensor<3x3xbf16>) -> tensor<3x3xbf16>
// CHECK:  return [[MUL]] : tensor<3x3xbf16>

module {

  func.func private @internal_dequantize_f32(
                           %input : tensor<*xf32>, %weight_scale : tensor<*xf32>) -> tensor<*xf32> {
    %mul = "tf.Mul"(%input, %weight_scale) : (tensor<*xf32>, tensor<*xf32>) -> tensor<*xf32>
    func.return %mul : tensor<*xf32>
  }

  func.func private @internal_conv3d_fn(
                         %input : tensor<*xf32>, %filter : tensor<*xi8>) -> tensor<*xf32> {