VSLDB $8, T1, T0, T0
	VSLDB $8, T2, T1, T1

	VACCQ  T0, RED1, CAR1
	VAQ    T0, RED1, T0
	VACCCQ T1, RED2, CAR1, CAR2
	VACQ   T1, RED2, CAR1, T1
	VAQ    T2, CAR2, T2

	// Last round
	VPERM T1, T0, SEL1, RED2    // d1 d0 d1 d0
	VPERM ZER, RED2, SEL2, RED1 // 0  d1 d0  0
	VSQ   RED1, RED2, RED2      // Guaranteed not to underflow

	VSLDB $8, T1, T0, T0
	VSLDB $8, T2, T1, T1

	VONE	V16
	VREPF	$0, V0, V1
index4loop:
	VL	(R7), V2           // load 16-bytes into V2
	VLL	R0, 16(R7), V3     // load 3-bytes into V3
	VSLDB	$1, V2, V3, V4     // V4=(V2:V3)<<1
	VSLDB	$2, V2, V3, V9     // V9=(V2:V3)<<1
	VSLDB	$3, V2, V3, V10    // V10=(V2:V3)<<1
	VCEQF	V1, V2, V5         // compare index 0, 4, ...
	VCEQF	V1, V4, V6         // compare index 1, 5, ...

	VPERM   ZER, RED2, SEL2, RED1 // 0  d1 d0  0
	VSUBUQM RED2, RED1, RED2      // VSQ   RED1, RED2, RED2      // Guaranteed not to underflow

	VSLDOI $8, T1, T0, T0 // VSLDB $8, T1, T0, T0
	VSLDOI $8, T2, T1, T1 // VSLDB $8, T2, T1, T1

	VADDCUQ  T0, RED1, CAR1       // VACCQ  T0, RED1, CAR1
	VADDUQM  T0, RED1, T0         // VAQ    T0, RED1, T0
	VADDECUQ T1, RED2, CAR1, CAR2 // VACCCQ T1, RED2, CAR1, CAR2

	"VPOPCT",
	"VREP",
	"VREPB",
	"VREPH",
	"VREPF",
	"VREPG",
	"VREPI",
	"VREPIB",
	"VREPIH",
	"VREPIF",
	"VREPIG",
	"VSCEF",
	"VSCEG",
	"VSEL",
	"VSL",
	"VSLB",
	"VSLDB",
	"VSRA",
	"VSRAB",
	"VSRL",
	"VSRLB",
	"VSEG",
	"VSEGB",
	"VSEGH",
	"VSEGF",
	"VST",
	"VSTEH",
	"VSTEF",
	"VSTEG",
	"VSTEB",
	"VSTM",
	"VSTL",
	"VSTRC",

	VLEIB	$15, $255, V0           // e70000fff040
	VLEIH	$7, $-32768, V15        // e7f080007041
	VLEIF	$2, $-43, V16           // e700ffd52843
	VLEIG	$1, $32767, V31         // e7f07fff1842
	VSLDB	$3, V1, V16, V18        // e72100030a77
	VERIMB	$2, V31, V1, V2         // e72f10020472
	VSEL	V1, V2, V3, V4          // e7412000308d
	VGFMAH	V21, V31, V24, V0       // e705f10087bc