SLD    $24, R21
	MTVSRD R21, V10
	VSPLTW $1, V10, V29
	VSLDOI $2, V29, V29, V30 // Mask 0x0000ff000000ff00...
	MOVD   $0xffff, R21
	SLD    $16, R21
	MTVSRD R21, V10
	VSPLTW $1, V10, V31      // Mask 0xffff0000ffff0000...
	VSPLTW $0, V0, V1        // Splat 1st word of separator

index4loop:
	VLOADSWAP(R7, R0, V2, V2)   // Load 16 bytes @R7 into V2

next4:
	VSPLTISB $0, V10            // Clear

	CMP     R8, R4
	BGE     loopexit        // Already at end?

	// vectorize if len(z) is >=3, else jump to scalar loop
	CMP     R4, $3
	BLT     scalar
	MTVSRD  R9, VS38        // s
	VSPLTB  $7, V6, V4
	MTVSRD  R5, VS39        // ŝ
	VSPLTB  $7, V7, V2
	ADD     $-2, R4, R16
	PCALIGN $16
loopback:
	ADD     $-1, R8, R10
	SLD     $3, R10
	LXVD2X  (R6)(R10), VS32 // load x[i-1], x[i]
	SLD     $3, R8, R12

	MOVD   $0x30, R10
	LXVD2X (HTBL)(R0), VH // Load H

	VSPLTISB $-16, XC2           // 0xf0
	VSPLTISB $1, T0              // one
	VADDUBM  XC2, XC2, XC2       // 0xe0
	VXOR     ZERO, ZERO, ZERO
	VOR      XC2, T0, XC2        // 0xe1
	VSLDOI   $15, XC2, ZERO, XC2 // 0xe1...
	VSLDOI   $1, ZERO, T0, T1    // ...1
	VADDUBM  XC2, XC2, XC2       // 0xc2...
	VSPLTISB $7, T2
	VOR      XC2, T1, XC2        // 0xc2....01

	VXOR	IN0, TMP, IN0       // vxor 1,1,6
	VSLDOI	$12, ZERO, TMP, TMP // vsldoi 6,0,6,12
	VXOR	IN0, TMP, IN0       // vxor 1,1,6

	VSLDOI	$8, ZERO, IN1, STAGE  // vsldoi 7,0,2,8
	VSPLTW	$3, IN0, TMP          // vspltw 6,1,3
	VXOR	TMP, IN1, TMP         // vxor 6,6,2
	VSLDOI	$12, ZERO, IN1, IN1   // vsldoi 2,0,2,12
	VADDUWM	RCON, RCON, RCON      // vadduwm 4,4,4
	VXOR	IN1, TMP, IN1         // vxor 2,2,6

	AVCMPNEZB
	AVCMPNEZBCC
	AVCMPNEB
	AVCMPNEBCC
	AVCMPNEH
	AVCMPNEHCC
	AVCMPNEW
	AVCMPNEWCC
	AVPERM
	AVPERMXOR
	AVPERMR
	AVBPERMQ
	AVBPERMD
	AVSEL
	AVSPLTB
	AVSPLTH
	AVSPLTW
	AVSPLTISB
	AVSPLTISH
	AVSPLTISW
	AVCIPH
	AVCIPHER
	AVCIPHERLAST
	AVNCIPH
	AVNCIPHER
	AVNCIPHERLAST
	AVSBOX
	AVSHASIGMA
	AVSHASIGMAW
	AVSHASIGMAD
	AVMRGEW

    private VisualCppInstall buildVisualCppInstall(String name, File vsPath, File basePath, VersionNumber version, Compatibility compatibility) {
        switch (compatibility) {
            case LEGACY:
                return buildLegacyVisualCppInstall(name, vsPath, basePath, version);
            case VS2017_OR_LATER:
                return buildVisualCppInstall(name, vsPath, basePath, version);
            default:

	SLD	$6, LEN
	ADD	INP, LEN, END

	CMP	INP, END
	BEQ	end

	MOVD	$·kcon(SB), TBL_STRT
	MOVD	$0x10, R_x010

#ifdef GOARCH_ppc64le
	MOVWZ	$8, TEMP
	LVSL	(TEMP)(R0), LEMASK
	VSPLTISB	$0x0F, KI
	VXOR	KI, LEMASK, LEMASK
#endif

	LXVW4X	(CTX)(R_x000), V0
	LXVW4X	(CTX)(R_x010), V4

	// unpack the input values into vector registers
	VSLDOI	$4, V0, V0, V1
	VSLDOI	$8, V0, V0, V2

	SLD	$6, LEN
	ADD	INP, LEN, END

	CMP	INP, END
	BEQ	end

	MOVD	$·kcon(SB), TBL_STRT
	MOVD	$0x10, R_x010

#ifdef GOARCH_ppc64le
	MOVWZ	$8, TEMP
	LVSL	(TEMP)(R0), LEMASK
	VSPLTISB	$0x0F, KI
	VXOR	KI, LEMASK, LEMASK
#endif

	LXVW4X	(CTX)(R_x000), V0
	LXVW4X	(CTX)(R_x010), V4

	// unpack the input values into vector registers
	VSLDOI	$4, V0, V0, V1
	VSLDOI	$8, V0, V0, V2

	MOVD	$0x0e0, R_x0e0
	MOVD	$0x0f0, R_x0f0
	MOVD	$0x100, R_x100
	MOVD	$0x110, R_x110


#ifdef GOARCH_ppc64le
	// Generate the mask used with VPERM for LE
	MOVWZ	$8, TEMP
	LVSL	(TEMP)(R0), LEMASK
	VSPLTISB	$0x0F, KI
	VXOR	KI, LEMASK, LEMASK
#endif

	LXVD2X	(CTX)(R_x000), VS32	// v0 = vs32
	LXVD2X	(CTX)(R_x010), VS34	// v2 = vs34
	LXVD2X	(CTX)(R_x020), VS36	// v4 = vs36

	BR	startbarConst
barcstTable:
	MOVD    $·CastBarConst(SB),R3

startbarConst:
	LVX	(R3),const1
	LVX	(R3+off16),const2

	VSLDOI	$8,V0,V0,V1
	VXOR	V0,V1,V0

#ifdef REFLECT
	VSPLTISB $1,V1
	VSL	V0,V1,V0
#endif

	VAND	V0,mask_64bit,V0

#ifndef	REFLECT

	VPMSUMD	V0,const1,V1
	VSLDOI	$8,zeroes,V1,V1
	VPMSUMD	V1,const2,V1
	VXOR	V0,V1,V0
	VSLDOI	$8,V0,zeroes,V0