x3 := byteorder.LeUint32(q[4*0x3:])
		x4 := byteorder.LeUint32(q[4*0x4:])
		x5 := byteorder.LeUint32(q[4*0x5:])
		x6 := byteorder.LeUint32(q[4*0x6:])
		x7 := byteorder.LeUint32(q[4*0x7:])
		x8 := byteorder.LeUint32(q[4*0x8:])
		x9 := byteorder.LeUint32(q[4*0x9:])
		xa := byteorder.LeUint32(q[4*0xa:])
		xb := byteorder.LeUint32(q[4*0xb:])
		xc := byteorder.LeUint32(q[4*0xc:])
		xd := byteorder.LeUint32(q[4*0xd:])

	// scramble 3 times
	AESENC	X8, X8
	AESENC	X9, X9
	AESENC	X10, X10
	AESENC	X11, X11
	AESENC	X12, X12
	AESENC	X13, X13
	AESENC	X14, X14
	AESENC	X15, X15

	AESENC	X8, X8
	AESENC	X9, X9
	AESENC	X10, X10
	AESENC	X11, X11
	AESENC	X12, X12
	AESENC	X13, X13
	AESENC	X14, X14
	AESENC	X15, X15

	AESENC	X8, X8
	AESENC	X9, X9
	AESENC	X10, X10

		reg: regInfo{
			inputs: []inputInfo{
				{0, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
				{1, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
				{2, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
			},
			outputs: []outputInfo{
				{0, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
			},
		},
	},
	{
		name:         "MINSD",

	// seq 1 206 | sed 's/.*/	z = x&/'
	var x1 [10<<20]byte
	var x2 [10<<20]byte
	var x3 [10<<20]byte
	var x4 [10<<20]byte
	var x5 [10<<20]byte
	var x6 [10<<20]byte
	var x7 [10<<20]byte
	var x8 [10<<20]byte
	var x9 [10<<20]byte
	var x10 [10<<20]byte
	var x11 [10<<20]byte
	var x12 [10<<20]byte
	var x13 [10<<20]byte
	var x14 [10<<20]byte
	var x15 [10<<20]byte
	var x16 [10<<20]byte
	var x17 [10<<20]byte

((Rsh64U|Lsh64)x32 <t> x y)  => (ISEL [0] (S(R|L)D <t> x y) (MOVDconst [0])        (CMPWUconst y [64]))
((Rsh64U|Lsh64)x16 <t> x y)  => (ISEL [2] (S(R|L)D <t> x y) (MOVDconst [0])        (CMPconst [0] (ANDconst [0xFFC0] y)))
((Rsh64U|Lsh64)x8  <t> x y)  => (ISEL [2] (S(R|L)D <t> x y) (MOVDconst [0])        (CMPconst [0] (ANDconst [0x00C0] y)))
(Rsh64x(64|32)     <t> x y)  => (ISEL [0] (SRAD    <t> x y) (SRADconst <t> x [63]) (CMP(U|WU)const y [64]))

	var x3 C.jthrowable = 0
	_ = x3
	var x4 C.jstring = 0
	_ = x4
	var x5 C.jarray = 0
	_ = x5
	var x6 C.jbooleanArray = 0
	_ = x6
	var x7 C.jbyteArray = 0
	_ = x7
	var x8 C.jcharArray = 0
	_ = x8
	var x9 C.jshortArray = 0
	_ = x9
	var x10 C.jintArray = 0
	_ = x10
	var x11 C.jlongArray = 0
	_ = x11
	var x12 C.jfloatArray = 0
	_ = x12
	var x13 C.jdoubleArray = 0

  // Build the shape of the refined type, if lhs is unranked it
  // will be directly the shape of the refined type, otherwise we merged by
  // taking the most specialized. This combines `10x?x?` and `?x?x8` into
  // `10x?x8`.
  if (!lhs_shape_type.hasRank()) {
    if (rhs_shape_type.hasRank()) {
      shape.append(rhs_shape_type.getShape().begin(),
                   rhs_shape_type.getShape().end());
      refined_shape = true;

// left shift
(Lsh(64|32|16|8)x64 <t> x y) && shiftIsBounded(v) => (SLL <t> x y)
(Lsh(64|32|16|8)x32 <t> x y) && shiftIsBounded(v) => (SLL <t> x y)
(Lsh(64|32|16|8)x16 <t> x y) && shiftIsBounded(v) => (SLL <t> x y)
(Lsh(64|32|16|8)x8  <t> x y) && shiftIsBounded(v) => (SLL <t> x y)

// signed right shift
(Rsh64x(64|32|16|8) <t> x y) && shiftIsBounded(v) => (SRA <t> x y)
(Rsh32x(64|32|16|8) <t> x y) && shiftIsBounded(v) => (SRA <t> (SignExt32to64 x) y)

ĦüA {3Åý »á6 ô Òb\9ºŽÞ쳬ô¹,w…BÉé)uÌÊŠ ù?Z \œáýóLjä‹h óÌ îBM†,êó(†ayÇ °páËʸ[É Ÿ® vâ鱟§î= °A 6Ž?¨ È'uBAL ódb%bn e½ 1 ×¥³n¤f‚Y!‘~ï‰+ ÝïAëˆÎØyþú==Âz_æ~? c{|»pÛ¼ å7SB‹ÒåLZ0eö+]51´Zmză§îVû·hÏÓ p¡sæ× ,°†ÝI »9l‡:‚ã¾öÈN· 2°k² Ô0¤û5\³mW[mâ? ŽCܤ«‡¡j( {7T` X8 sbA'éY‹P 5z˜ù·|4,þ:KWˆø˜˜ò |°Vd?<õ‹ aíêïõucÄ]º±ÊƒË,h$ b½`ØŠö´q¡©ç 8d .ZŒ‘ÿØ äVÎÉ ¤ jü3nº ܯ†¿| ¿O‡† ÂJŸä |ÆoVˆÐØ’œ˜Ä( ’Feƒ¢eÜô®ÊgM3YhC Ê×&$ öÞ Ê¢$‡~´ ¬Ü¡:ñqø”÷Zs¤â­¿,×ÿè2 >=þÞ“^´/A@cÐü…×| â äÄ ¢Îx„o:vnðEc‰º )Ö‹LgöX€Òºƒ7½ ‰}e€ÈÓ""­5‘–©@Ü Ì†@×¹f 53Å)÷T…öš[Ú=8®S#´]™9Ÿç%...