// Checks if the value has only one user.
def HasOneUse : Constraint<CPred<"$0.hasOneUse()">>;

// If we see a Conv2D op followed by Mul, then multiply the filter
// with the value in Mul.
def FuseMulAndConv2D :
  Pat<(TF_MulOp:$mul (TF_Conv2DOp:$conv $input,
                          (Arith_ConstantOp:$filter F32ElementsAttr:$filter_value),

	z.Square(x)
	z.Mul(x, z)
	t0.Square(z)
	for s := 1; s < 2; s++ {
		t0.Square(t0)
	}
	z.Mul(z, t0)
	t0.Square(z)
	for s := 1; s < 4; s++ {
		t0.Square(t0)
	}
	z.Mul(z, t0)
	t0.Square(z)
	for s := 1; s < 8; s++ {
		t0.Square(t0)
	}
	z.Mul(z, t0)
	t0.Square(z)
	for s := 1; s < 16; s++ {
		t0.Square(t0)
	}
	z.Mul(z, t0)
	for s := 0; s < 32; s++ {

	z.Square(x)
	z.Mul(x, z)
	z.Square(z)
	t1.Mul(x, z)
	z.Square(t1)
	for s := 1; s < 3; s++ {
		z.Square(z)
	}
	z.Mul(t1, z)
	t0.Square(z)
	for s := 1; s < 6; s++ {
		t0.Square(t0)
	}
	t0.Mul(z, t0)
	t2.Square(t0)
	for s := 1; s < 12; s++ {
		t2.Square(t2)
	}
	t0.Mul(t0, t2)
	for s := 0; s < 6; s++ {
		t0.Square(t0)
	}
	z.Mul(z, t0)
	t0.Square(z)
	t2.Mul(x, t0)

	}
	z.Mul(z, t0)
	t0.Square(z)
	t0.Mul(x, t0)
	for s := 0; s < 64; s++ {
		t0.Square(t0)
	}
	z.Mul(z, t0)
	t0.Square(z)
	t0.Mul(x, t0)
	for s := 0; s < 129; s++ {
		t0.Square(t0)
	}
	z.Mul(z, t0)
	t0.Square(z)
	t0.Mul(x, t0)
	for s := 0; s < 259; s++ {
		t0.Square(t0)
	}
	z.Mul(z, t0)
	for s := 0; s < 2; s++ {
		z.Square(z)
	}
	z.Mul(x, z)

	return e.Set(z)

	// 386:"SHLL\t[$]5",-"IMULL"
	// arm:"SLL\t[$]5",-"MUL"
	// arm64:"LSL\t[$]5",-"MUL"
	// ppc64x:"SLD\t[$]5",-"MUL"
	a := n1 * 32

	// amd64:"SHLQ\t[$]6",-"IMULQ"
	// 386:"SHLL\t[$]6",-"IMULL"
	// arm:"SLL\t[$]6",-"MUL"
	// arm64:`NEG\sR[0-9]+<<6,\sR[0-9]+`,-`LSL`,-`MUL`
	// ppc64x:"SLD\t[$]6","NEG\\sR[0-9]+,\\sR[0-9]+",-"MUL"
	b := -64 * n2

	return a, b
}

func Mul_96(n int) int {

	var t2 = new(fiat.P384Element)

	z.Square(x)
	z.Mul(x, z)
	z.Square(z)
	t0.Mul(x, z)
	z.Square(t0)
	for s := 1; s < 3; s++ {
		z.Square(z)
	}
	t1.Mul(t0, z)
	t2.Square(t1)
	z.Mul(x, t2)
	for s := 0; s < 5; s++ {
		t2.Square(t2)
	}
	t1.Mul(t1, t2)
	t2.Square(t1)
	for s := 1; s < 12; s++ {
		t2.Square(t2)
	}
	t1.Mul(t1, t2)
	for s := 0; s < 7; s++ {
		t1.Square(t1)
	}

	var t1 = new(fiat.P224Element)

	r.Square(x)
	r.Mul(x, r)
	r.Square(r)
	r.Mul(x, r)
	t0.Square(r)
	for s := 1; s < 3; s++ {
		t0.Square(t0)
	}
	t0.Mul(r, t0)
	t1.Square(t0)
	r.Mul(x, t1)
	for s := 0; s < 5; s++ {
		t1.Square(t1)
	}
	t0.Mul(t0, t1)
	t1.Square(t0)
	for s := 1; s < 12; s++ {
		t1.Square(t1)
	}
	t0.Mul(t0, t1)
	t1.Square(t0)
	for s := 1; s < 7; s++ {
		t1.Square(t1)

	}
	i := new(big.Int).Lsh(h, 1)
	i.Mul(i, i)
	j := new(big.Int).Mul(h, i)

	s1 := new(big.Int).Mul(y1, z2)
	s1.Mul(s1, z2z2)
	s1.Mod(s1, curve.P)
	s2 := new(big.Int).Mul(y2, z1)
	s2.Mul(s2, z1z1)
	s2.Mod(s2, curve.P)
	r := new(big.Int).Sub(s2, s1)
	if r.Sign() == -1 {
		r.Add(r, curve.P)
	}
	yEqual := r.Sign() == 0
	if xEqual && yEqual {
		return curve.doubleJacobian(x1, y1, z1)
	}

	MUL	x2, x2, t0
	ADCS	t0, acc4, acc4
	UMULH	x2, x2, t1
	ADCS	t1, acc5, acc5

	MUL	x3, x3, t0
	ADCS	t0, acc6, acc6
	UMULH	x3, x3, t1
	ADC	t1, acc7, acc7
	// First reduction step
	MUL	acc0, hlp1, hlp0

	MUL	const0, hlp1, t0
	ADDS	t0, acc0, acc0
	UMULH	const0, hlp0, t1

	MUL	const1, hlp0, t0
	ADCS	t0, acc1, acc1
	UMULH	const1, hlp0, y0

	MUL	const2, hlp0, t0
	ADCS	t0, acc2, acc2

  input: "Case/Case/input_0/_7"
  input: "Case/branch0/_0/mul/y"
  attr {
    key: "T"
    value {
      type: DT_FLOAT
    }
  }
}
node {
  name: "Case/branch1/_1/mul_0"
  op: "Mul"
  input: "Case/Case/input_0/_7:1"
  input: "Case/branch1/_1/mul/y"
  attr {
    key: "T"
    value {
      type: DT_FLOAT
    }
  }
}
node {
  name: "Case/merge/_9"
  op: "Merge"