BGEU	X5, X6, 2(PC)				// 63f46200

	// 2.6: Load and Store Instructions
	LW	(X5), X6				// 03a30200
	LW	4(X5), X6				// 03a34200
	LWU	(X5), X6				// 03e30200
	LWU	4(X5), X6				// 03e34200
	LH	(X5), X6				// 03930200
	LH	4(X5), X6				// 03934200
	LHU	(X5), X6				// 03d30200
	LHU	4(X5), X6				// 03d34200
	LB	(X5), X6				// 03830200
	LB	4(X5), X6				// 03834200
	LBU	(X5), X6				// 03c30200

	MOV	b+8(FP), X6
	MOV	$1, X7
	BGEU	X5, X6, b
	MOV	$0, X7
b:
	MOV	X7, r+16(FP)
	RET

// func testBGEZ(a int64) (r bool)
TEXT ·testBGEZ(SB),NOSPLIT,$0-9
	MOV	a+0(FP), X5
	MOV	$1, X6
	BGEZ	X5, b
	MOV	$0, X6
b:
	MOV	X6, r+8(FP)
	RET

// func testBGT(a, b int64) (r bool)
TEXT ·testBGT(SB),NOSPLIT,$0-17
	MOV	a+0(FP), X5
	MOV	b+8(FP), X6
	MOV	$1, X7
	BGT	X5, X6, b
	MOV	$0, X7
b:

	RORIW	$32, X5, X6			// ERROR "immediate out of range 0 to 31"
	SLLIW	$32, X5, X6			// ERROR "immediate out of range 0 to 31"
	SRLIW	$32, X5, X6			// ERROR "immediate out of range 0 to 31"
	SRAIW	$32, X5, X6			// ERROR "immediate out of range 0 to 31"
	RORIW	$-1, X5, X6			// ERROR "immediate out of range 0 to 31"
	SLLIW	$-1, X5, X6			// ERROR "immediate out of range 0 to 31"

cmp1a:
	SLTU	X9, X8, X5
	SLTU	X8, X9, X6
	JMP	cmp_ret
cmp1b:
	SLTU	X16, X15, X5
	SLTU	X15, X16, X6
	JMP	cmp_ret
cmp1c:
	SLTU	X18, X17, X5
	SLTU	X17, X18, X6
	JMP	cmp_ret
cmp1d:
	SLTU	X20, X19, X5
	SLTU	X19, X20, X6
	JMP	cmp_ret
cmp1e:
	SLTU	X22, X21, X5
	SLTU	X21, X22, X6
	JMP	cmp_ret
cmp1f:
	SLTU	X24, X23, X5
	SLTU	X23, X24, X6
	JMP	cmp_ret
cmp1g:

	VRSQRT28SD X8, X6, K1, X6                          // 62d2cd09cdf0 or 62d2cd29cdf0 or 62d2cd49cdf0
	VRSQRT28SD X6, X6, K1, X6                          // 62f2cd09cdf6 or 62f2cd29cdf6 or 62f2cd49cdf6
	VRSQRT28SD X0, X6, K1, X6                          // 62f2cd09cdf0 or 62f2cd29cdf0 or 62f2cd49cdf0

	MOV	(((index-16)&0xf)*8)(X19), X21; \
	ROR	$19, X5, X7; \
	ROR	$61, X5, X8; \
	SRL	$6, X5; \
	XOR	X7, X5; \
	XOR	X8, X5; \
	ADD	X9, X5; \
	ROR	$1, X6, X7; \
	ROR	$8, X6, X8; \
	SRL	$7, X6; \
	XOR	X7, X6; \
	XOR	X8, X6; \
	ADD	X6, X5; \
	ADD	X21, X5; \
	MOV	X5, ((index&0xf)*8)(X19)

// Calculate T1 in X5.
// h is also used as an accumulator. Wt is passed in X5.
//   T1 = h + BIGSIGMA1(e) + Ch(e, f, g) + Kt + Wt

	/* whether x is NaN */; \
	FEQD	F0, F0, X6;	\
	BNEZ	X6, 3(PC);	\
	/* return NaN if x is NaN */; \
	MOVD	F0, ret+8(FP); 	\
	RET;			\
	MOV	$PosInf, X6;	\
	FMVDX	X6, F1;		\
	FABSD	F0, F2;		\
	/* if abs(x) > +Inf, return Inf instead of round(x) */; \
	FLTD	F1, F2, X6;	\
	/* Inf should keep same signed with x then return */;	\
	BEQZ	X6, 3(PC); \
	FCVTLD.MODE	F0, X6;	\
	FCVTDL	X6, F1;		\

	// t1 := s2 * (L1 + s4*(L3+s4*(L5+s4*L7)))
	MOVSD   $L7, X6
	MULSD   X5, X6
	ADDSD   $L5, X6
	MULSD   X5, X6
	ADDSD   $L3, X6
	MULSD   X5, X6
	ADDSD   $L1, X6
	MULSD   X6, X4 // x1= k, x2= f, x3= s, x4= t1, x5= s4
	// t2 := s4 * (L2 + s4*(L4+s4*L6))
	MOVSD   $L6, X6
	MULSD   X5, X6
	ADDSD   $L4, X6
	MULSD   X5, X6
	ADDSD   $L2, X6
	MULSD   X6, X5 // x1= k, x2= f, x3= s, x4= t1, x5= t2
	// R := t1 + t2

	asm := `
TEXT _stub(SB),$0-0
	LB	4096(X5), X6
	LH	4096(X5), X6
	LW	4096(X5), X6
	LD	4096(X5), X6
	LBU	4096(X5), X6
	LHU	4096(X5), X6
	LWU	4096(X5), X6
	SB	X6, 4096(X5)
	SH	X6, 4096(X5)
	SW	X6, 4096(X5)
	SD	X6, 4096(X5)

	FLW	4096(X5), F6
	FLD	4096(X5), F6
	FSW	F6, 4096(X5)
	FSD	F6, 4096(X5)

	MOVB	4096(X5), X6
	MOVH	4096(X5), X6
	MOVW	4096(X5), X6
	MOV	4096(X5), X6

	ADD	X9, X22, X29	// next c

	MULHU	X11, X6, X12	// z_hi[1] = x[1] * y
	MUL	X11, X6, X11	// z_lo[1] = x[1] * y
	ADD	X11, X13, X21	// z_lo[1] = x[1] * y + z[1]
	SLTU	X11, X21, X22
	ADD	X12, X22, X12	// z_hi[1] = x[1] * y + z[1]
	ADD	X21, X29, X13	// z_lo[1] = x[1] * y + z[1] + c
	SLTU	X21, X13, X22
	ADD	X12, X22, X29	// next c

	MULHU	X14, X6, X15	// z_hi[2] = x[2] * y
	MUL	X14, X6, X14	// z_lo[2] = x[2] * y