VSHASIGMAD $2, $15, V1, V2      // 104196c2

	LXVD2X (R3)(R4), VS1            // 7c241e98
	LXVD2X (R3)(R0), VS1            // 7c201e98
	LXVD2X (R3), VS1                // 7c201e98
	LXVDSX (R3)(R4), VS1            // 7c241a98
	LXVDSX (R3)(R0), VS1            // 7c201a98
	LXVDSX (R3), VS1                // 7c201a98
	LXVH8X (R3)(R4), VS1            // 7c241e58
	LXVH8X (R3)(R0), VS1            // 7c201e58

	// cond is R1 + 24 (cond offset) + 32
	LXVDSX (R1)(R21), SEL
	VSPLTISB $0, ZER
	// SEL controls whether to store a or b
	VCMPEQUD SEL, ZER, SEL

	LXVD2X (P1ptr+R0), X1H
	LXVD2X (P1ptr+R16), X1L
	LXVD2X (P1ptr+R17), Y1H
	LXVD2X (P1ptr+R18), Y1L
	LXVD2X (P1ptr+R19), Z1H
	LXVD2X (P1ptr+R20), Z1L

	LXVD2X (P2ptr+R0), X2H
	LXVD2X (P2ptr+R16), X2L
	LXVD2X (P2ptr+R17), Y2H
	LXVD2X (P2ptr+R18), Y2L

		// number of moves are generated based on the
		// size.
		// When moving >= 64 bytes a loop is used
		//	MOVD len/32,REG_TMP
		//	MOVD REG_TMP,CTR
		//	MOVD $16,REG_TMP
		// top:
		//	LXVD2X (R0)(R21),VS32
		//	LXVD2X (R31)(R21),VS33
		//	ADD $32,R21
		//	STXVD2X VS32,(R0)(R20)
		//	STXVD2X VS33,(R31)(R20)
		//	ADD $32,R20
		//	BC 16,0,top
		// Bytes not moved by this loop are moved