feMul(v, x, y)
	return v
}

// Square sets v = x * x, and returns v.
func (v *Element) Square(x *Element) *Element {
	feSquare(v, x)
	return v
}

// Mult32 sets v = x * y, and returns v.
func (v *Element) Mult32(x *Element, y uint32) *Element {
	x0lo, x0hi := mul51(x.l0, y)
	x1lo, x1hi := mul51(x.l1, y)
	x2lo, x2hi := mul51(x.l2, y)
	x3lo, x3hi := mul51(x.l3, y)
	x4lo, x4hi := mul51(x.l4, y)

		t.Errorf("Swap failed")
	}
}

func TestMult32(t *testing.T) {
	mult32EquivalentToMul := func(x Element, y uint32) bool {
		t1 := new(Element)
		for i := 0; i < 100; i++ {
			t1.Mult32(&x, y)
		}

		ty := new(Element)
		ty.l0 = uint64(y)

		t2 := new(Element)
		for i := 0; i < 100; i++ {
			t2.Multiply(&x, ty)
		}

					test.in, out32, err, test.out, test.err, out)
			}
		}
	}
	for _, test := range atof32tests {
		out, err := ParseFloat(test.in, 32)
		out32 := float32(out)
		if float64(out32) != out {
			t.Errorf("ParseFloat(%v, 32) = %v, not a float32 (closest is %v)", test.in, out, float64(out32))
			continue
		}
		outs := FormatFloat(float64(out32), 'g', -1, 32)
		if outs != test.out || !reflect.DeepEqual(err, test.err) {

	// 	x1:x0 := r.Uint64()
	// 	0:hi, lo1:lo0 := bits.Mul64(x1:x0, 0:n)
	// Writing out the multiplication in terms of bits.Mul32 allows
	// using direct hardware instructions and avoiding
	// the computations involving these zeros.
	x := r.Uint64()
	lo1a, lo0 := bits.Mul32(uint32(x), n)
	hi, lo1b := bits.Mul32(uint32(x>>32), n)
	lo1, c := bits.Add32(lo1a, lo1b, 0)
	hi += c
	if lo1 == 0 && lo0 < uint32(n) {
		n64 := uint64(n)

		(Select0 <typ.UInt32> (Sub32carry (Int64Lo x) (Int64Lo y))))

(Mul64 x y) =>
	(Int64Make
		(Add32 <typ.UInt32>
			(Mul32 <typ.UInt32> (Int64Lo x) (Int64Hi y))
			(Add32 <typ.UInt32>
				(Mul32 <typ.UInt32> (Int64Hi x) (Int64Lo y))
				(Select0 <typ.UInt32> (Mul32uhilo (Int64Lo x) (Int64Lo y)))))
		(Select1 <typ.UInt32> (Mul32uhilo (Int64Lo x) (Int64Lo y))))

(And64 x y) =>

//
// This function's execution time does not depend on the inputs.
func Mul(x, y uint) (hi, lo uint) {
	if UintSize == 32 {
		h, l := Mul32(uint32(x), uint32(y))
		return uint(h), uint(l)
	}
	h, l := Mul64(uint64(x), uint64(y))
	return uint(h), uint(l)
}

// Mul32 returns the 64-bit product of x and y: (hi, lo) = x * y
// with the product bits' upper half returned in hi and the lower
// half returned in lo.
//

		x, y      uint32
		hi, lo, r uint32
	}{
		{1 << 31, 2, 1, 0, 1},
		{0xc47dfa8c, 50911, 0x98a4, 0x998587f4, 13},
		{_M32, _M32, _M32 - 1, 1, 42},
	} {
		testMul("Mul32", Mul32, a.x, a.y, a.hi, a.lo)
		testMul("Mul32 symmetric", Mul32, a.y, a.x, a.hi, a.lo)
		testDiv("Div32", Div32, a.hi, a.lo+a.r, a.y, a.x, a.r)
		testDiv("Div32 symmetric", Div32, a.hi, a.lo+a.r, a.x, a.y, a.r)
	}
}

	if _W >= 32 {
		runModWTests(t, modWTests32)
	}
	if _W >= 64 {
		runModWTests(t, modWTests64)
	}
}

var montgomeryTests = []struct {
	x, y, m      string
	k0           uint64
	out32, out64 string
}{
	{
		"0xffffffffffffffffffffffffffffffffffffffffffffffffe",
		"0xffffffffffffffffffffffffffffffffffffffffffffffffe",
		"0xfffffffffffffffffffffffffffffffffffffffffffffffff",
		1,

pkg math/bits, func Div32(uint32, uint32, uint32) (uint32, uint32)
pkg math/bits, func Div64(uint64, uint64, uint64) (uint64, uint64)
pkg math/bits, func Mul(uint, uint) (uint, uint)
pkg math/bits, func Mul32(uint32, uint32) (uint32, uint32)
pkg math/bits, func Mul64(uint64, uint64) (uint64, uint64)
pkg math/bits, func Sub(uint, uint, uint) (uint, uint)
pkg math/bits, func Sub32(uint32, uint32, uint32) (uint32, uint32)

	{name: "Sub32F", argLength: 2},
	{name: "Sub64F", argLength: 2},

	{name: "Mul8", argLength: 2, commutative: true}, // arg0 * arg1
	{name: "Mul16", argLength: 2, commutative: true},
	{name: "Mul32", argLength: 2, commutative: true},
	{name: "Mul64", argLength: 2, commutative: true},
	{name: "Mul32F", argLength: 2, commutative: true},
	{name: "Mul64F", argLength: 2, commutative: true},