t.Errorf("#%d got %s want %s", i, z.utoa(10), out.utoa(10))
		}
	}
}

func FuzzExpMont(f *testing.F) {
	f.Fuzz(func(t *testing.T, x1, x2, x3, y1, y2, y3, m1, m2, m3 uint) {
		if m1 == 0 && m2 == 0 && m3 == 0 {
			return
		}
		x := new(Int).SetBits([]Word{Word(x1), Word(x2), Word(x3)})
		y := new(Int).SetBits([]Word{Word(y1), Word(y2), Word(y3)})
		m := new(Int).SetBits([]Word{Word(m1), Word(m2), Word(m3)})

				st.MakePod().Name("p-a4").Node("node-a").Label("foo", "").Obj(),
				st.MakePod().Name("p-b1").Node("node-b").Label("foo", "").Obj(),
				st.MakePod().Name("p-b2").Node("node-b").Label("foo", "").Obj(),
				st.MakePod().Name("p-x1").Node("node-x").Label("foo", "").Obj(),
				st.MakePod().Name("p-y1").Node("node-y").Label("foo", "").Obj(),
				st.MakePod().Name("p-y2").Node("node-y").Label("foo", "").Obj(),

	//	https://learn.microsoft.com/en-us/cpp/build/x64-calling-convention?view=msvc-170
_4args:
	MOVQ	24(SI), R9
	MOVQ	R9, X3
_3args:
	MOVQ	16(SI), R8
	MOVQ	R8, X2
_2args:
	MOVQ	8(SI), DX
	MOVQ	DX, X1
_1args:
	MOVQ	0(SI), CX
	MOVQ	CX, X0
_0args:

	// Call stdcall function.
	CALL	AX

	ADDQ	$(const_maxArgs*8), SP

	// Return result.
	MOVQ	0(SP), CX
	MOVQ	8(SP), SP
	MOVQ	AX, libcall_r1(CX)

}

func newMACGeneric(key *[32]byte) macGeneric {
	m := macGeneric{}
	initialize(key, &m.macState)
	return m
}

// macState holds numbers in saturated 64-bit little-endian limbs. That is,
// the value of [x0, x1, x2] is x[0] + x[1] * 2⁶⁴ + x[2] * 2¹²⁸.
type macState struct {
	// h is the main accumulator. It is to be interpreted modulo 2¹³⁰ - 5, but
	// can grow larger during and after rounds. It must, however, remain below

(9) CSEL, CSELW, CSNEG, CSNEGW, CSINC, CSINCW <cond>, <Rn>, <Rm>, <Rd> ;
FCSELD, FCSELS <cond>, <Fn>, <Fm>, <Fd>

Examples:

	CSEL GT, R0, R19, R1        <=>    csel x1, x0, x19, gt
	CSNEGW GT, R7, R17, R8      <=>    csneg w8, w7, w17, gt
	FCSELD EQ, F15, F18, F16    <=>    fcsel d16, d15, d18, eq

(10) TBNZ, TBZ $<imm>, <Rt>, <label>

from tensorflow.python.types import core
from tensorflow.python.util import tf_export

# A representative sample is a map of: input_key -> input_value.
# Ex.: {'dense_input': tf.constant([1, 2, 3])}
# Ex.: {'x1': np.ndarray([4, 5, 6]}
RepresentativeSample = Mapping[str, core.TensorLike]

# A representative dataset is an iterable of representative samples.
RepresentativeDataset = Iterable[RepresentativeSample]

	MOV	g, m_g0(X11)			// curm.g0 = g

	// switch to crashstack
	MOV	(g_stack+stack_hi)(g), X11
	SUB	$(4*8), X11
	MOV	X11, X2

	// call target function
	MOV	0(CTXT), X10
	JALR	X1, X10

	// should never return
	CALL	runtime·abort(SB)
	UNDEF

/*
 * support for morestack
 */

// Called during function prolog when more stack is needed.

	v3 := v1 + v1 + v1 ^ v1 | 3 + v1 ^ v1 | v1 ^ v1
	v1-- // dev.ssa doesn't see this one
	return v1 ^ v1*v1 - v3
}

func testSubConst(t *testing.T) {
	x1 := sub1_ssa()
	want1 := uint64(6)
	if x1 != want1 {
		t.Errorf("sub1_ssa()=%d, got %d", want1, x1)
	}
	x2 := sub2_ssa()
	want2 := uint8(251)
	if x2 != want2 {
		t.Errorf("sub2_ssa()=%d, got %d", want2, x2)
	}
}

//go:noinline

			x0 := p.getConstant(prog, op, &a[0])
			x1 := p.getConstant(prog, op, &a[1])
			x2 := int64(p.getRegister(prog, op, &a[2]))
			x3 := int64(p.getRegister(prog, op, &a[3]))
			x4 := int64(p.getRegister(prog, op, &a[4]))
			x5 := p.getConstant(prog, op, &a[5])
			// Cond is handled specially for this instruction.
			offset, MRC, ok := arch.ARMMRCOffset(op, cond, x0, x1, x2, x3, x4, x5)
			if !ok {

// dy_1/dx_m dy_2/dx_m ... dy_n/dx_m
//
// If x or y is complex, each complex entry is "expanded" into a real and
// imaginary entry, and the Jacobian is organized as above on the expanded list.
// e.g.
// [y1, y2] = Square([x1, x2]) where x and y are complex.
// Writing
// x = [x1_real, x1_imag, x2_real, x2_imag]
// y = [y1_real, y1_imag, y2_real, y2_imag]
// the Jacobian transpose is
// the 4x4 matrix: