MOV	3*8(X7), X17	// x[3]

	MULHU	X8, X6, X9	// z_hi[0] = x[0] * y
	MUL	X8, X6, X8	// z_lo[0] = x[0] * y
	ADD	X8, X10, X21	// z_lo[0] = x[0] * y + z[0]
	SLTU	X8, X21, X22
	ADD	X9, X22, X9	// z_hi[0] = x[0] * y + z[0]
	ADD	X21, X29, X10	// z_lo[0] = x[0] * y + z[0] + c
	SLTU	X21, X10, X22
	ADD	X9, X22, X29	// next c

	MULHU	X11, X6, X12	// z_hi[1] = x[1] * y
	MUL	X11, X6, X11	// z_lo[1] = x[1] * y

    %2 = "tf.MatMul"(%arg0, %1) {attr_map = "0:transpose_a,1:transpose_a", device = "", transpose_a = false, transpose_b = false} : (tensor<1x2x2x2xf32>, tensor<2x1024xf32>) -> tensor<*xf32>
    %3 = "tf.Mul"(%2, %cst_0) : (tensor<*xf32>, tensor<f32>) -> tensor<*xf32>
    return %3 : tensor<*xf32>
  }

// CHECK-LABEL: func @not_propagate_matmul

// CHECK:         %[[SUB:.*]] = s32[] subtract(s32[] %[[CONVERT_1]],
// CHECK:         %[[CONVERT_2:.*]] = f32[] convert(s32[] %[[SUB]])
// CHECK:         %[[MUL:.*]] = f32[] multiply(f32[] %[[CONVERT_2]],

  %1 = "tfl.mul"(%0, %cst_2) {fused_activation_function = "NONE"} : (tensor<1x256xbf16>, tensor<32x1x256xbf16>) -> tensor<32x1x256xbf16>
  func.return %1 : tensor<32x1x256xbf16>

// CHECK:  %[[V0:.*]] = "tfl.fully_connected"(%arg0, {{.*}}) <{{{.*}}}> : (tensor<1x10368xbf16>, tensor<256x10368xbf16>, none) -> tensor<1x256xbf16>

      %outputs_9, %control_10 = tf_executor.island wraps "tf.Mul"(%outputs_7, %outputs) {T = "tfdtype$DT_FLOAT", _tpu_replicate = "cluster", device = "", name = "mul"} : (tensor<0xf32>, tensor<f32>) -> tensor<0xf32>

			kInv = new(big.Int).ModInverse(k, N)

			r, _ = c.ScalarBaseMult(k.Bytes())
			r.Mod(r, N)
			if r.Sign() != 0 {
				break
			}
		}

		e := hashToInt(hash, c)
		s = new(big.Int).Mul(priv.D, r)
		s.Add(s, e)
		s.Mul(s, kInv)
		s.Mod(s, N) // N != 0
		if s.Sign() != 0 {
			break
		}
	}

	return encodeSignature(r.Bytes(), s.Bytes())
}

    %6 = "tf.Mul"(%arg2, %arg4) : (tensor<f32>, tensor<512xf32>) -> tensor<512xf32>
    %7 = "tf.Div"(%6, %arg6) : (tensor<512xf32>, tensor<f32>) -> tensor<512xf32>
    %8 = "tf.Cast"(%5) {Truncate = false} : (tensor<1x2240x1120x512xi32>) -> tensor<1x2240x1120x512xf32>
    %9 = "tf.Mul"(%7, %8) : (tensor<512xf32>, tensor<1x2240x1120x512xf32>) -> tensor<1x2240x1120x512xf32>

	}

	n := pub.Q.BitLen()
	if n%8 != 0 {
		return false
	}
	z := new(big.Int).SetBytes(hash)

	u1 := new(big.Int).Mul(z, w)
	u1.Mod(u1, pub.Q)
	u2 := w.Mul(r, w)
	u2.Mod(u2, pub.Q)
	v := u1.Exp(pub.G, u1, pub.P)
	u2.Exp(pub.Y, u2, pub.P)
	v.Mul(v, u2)
	v.Mod(v, pub.P)
	v.Mod(v, pub.Q)

	return v.Cmp(r) == 0

		FieldD string `gorm:"uniqueIndex;index"` // uniqueIndex and index

		FieldE1 string `gorm:"uniqueIndex:uniq_field_e1_e2"` // mul uniqueIndex
		FieldE2 string `gorm:"uniqueIndex:uniq_field_e1_e2"`

		FieldF1 string `gorm:"uniqueIndex:uniq_field_f1_f2;index"` // mul uniqueIndex and index
		FieldF2 string `gorm:"uniqueIndex:uniq_field_f1_f2;"`

		FieldG string `gorm:"unique;uniqueIndex"` // unique and uniqueIndex

// CHECK: %[[cast:.*]] = "tf.Cast"(%[[sub]]) <{Truncate = false}> : (tensor<1x3x2x2xi32>) -> tensor<1x3x2x2xf32>
// CHECK: %[[dequant1:.*]] = "tf.Mul"(%[[cast]]
// CHECK: %[[relu:.*]] = "tf.Relu"(%[[dequant1]]
// CHECK: %[[clamped:.*]] = "tf.Minimum"(%[[relu]]

// CHECK: %[[rescale1:.*]] = "tf.Mul"(%[[cast]]
// CHECK: %[[add2:.*]] = "tf.AddV2"(%[[rescale1]]
// CHECK: %[[maximum2:.*]] = "tf.Maximum"(%[[add2]]