MIPS64寄存器与指令集

 

MIPS64指令集

 

需要注意的与32位的区别(32为的w:word  ----- 64为的d: double)：

运算类：add变成dadd，同理sub/mul/div变成dsub/dmul/ddiv

访存: lw变成ld, sw变成sd

寄存器名： S0/S1…   全部变为r0/r1/… / r31

 

I/O映射表

 

 

在符号区域定义的符号的值是当前符号所对应的内存地址,使用的是字节地址

如:

.data

A :  .word 32

A的值并不是32，而是其所在的内存位置 0

 

在.text定义的变量记录的是当前的字地址

.text

Loop: dadd $s0,$s1,$s2

 

来源：网上资料

64位版本开头以“d”表示，无符号数以“u”结尾，立即数通常以“i”结尾，字节操作以“b”结尾，双字操作以“d”结尾，字操作以“w”结尾

序号	伪指令	功能
1	.data	start of data segment
2	.text	start of code segment
3	.code	start of code segment (same as .text)
4	.org	start address
5	.space	leave n empty bytes
6	.asciiz	enters zero terminated ascii string
7	.ascii	enter ascii string
8	.align	align to n-byte boundary
9	.word   ,..	enters word(s) of data (64-bits)
10	.byte   ,..	enter bytes
11	.word16 ,..	enters 16 bit number(s)
12	.word32 ,..	enters 32 bit number(s)
13	.double ,..	enters floating-point number(s)

 

序号	指令	功能
	访存
	lb reg,imm(reg)	load byte
	lbu reg,imm(reg)	load byte unsigned
	lh reg,imm(reg)	load 16-bit half-word
	lhu reg,imm(reg)	load 16-bit half word unsigned
	lw reg,imm(reg)	load 32-bit word
	lwu reg,imm(reg)	load 32-bit word unsigned
	ld reg,imm(reg)	load 64-bit double-word
	l.s freg,imm(reg)	load 32-bit floating-point(single FP )
	l.d freg,imm(reg)	load 64-bit floating-point(double FP )
	sb reg,imm(reg)	store byte
	sh reg,imm(reg)	store 16-bit half-word
	sw reg,imm(reg)	store 32-bit word
	sd reg,imm(reg)	store 64-bit double-word
	s.s freg,imm(reg)	store 32-bit floating-point
	s.d freg,imm(reg)	store 64-bit floating-point
	传送
	movz r1，r2,r3	If(r3==0) r1=r2
	movn reg,reg,reg	move if register not equal to zero
	mov.s freg,freg	move floating-point(single FP )
	mov.d freg,freg	move floating-point(double FP )
	mtc1 reg,freg	move data from integer register to FP register
	mfc1 reg,freg	move data from FP register to integer register
	lui  reg,imm	load upper half of register immediate
	算术
	dadd reg,reg,reg	add integers
	daddu reg,reg,reg	add integers unsigned
	daddi reg,reg,imm	add immediate
	daddui reg,reg,imm	add immediate unsigned( WinMIPS64格式 )
	daddiu reg,reg,imm	add immediate unsigned( MIPSsim格式 )
	add.s freg,freg,freg	add floating-point(single FP )
	add.d freg,freg,freg	add floating-point(double FP )
	dsub reg,reg,reg	subtract integers
	dsubu reg,reg,reg	subtract integers unsigned
	sub.s freg,freg,freg	subtract floating-point(single FP )
	sub.d freg,freg,freg	subtract floating-point(double FP )
	dmul reg,reg,reg	signed integer multiplication
	dmulu reg,reg,reg	unsigned integer multiplication
	mul.d freg,freg,freg	multiply floating-point
	ddiv reg,reg,reg	signed integer division
	ddivu reg,reg,reg	unsigned integer division
	div.d freg,freg,freg	divide floating-point
	逻辑
	and reg,reg,reg	logical and
	andi reg,reg,imm	logical and immediate
	or reg,reg,reg	logical or
	ori reg,reg,imm	logical or immediate
	xor reg,reg,reg	logical xor
	xori reg,reg,imm	exclusive or immediate
	dsll reg,reg,imm	shift left logical
	dsllv reg,reg,reg	shift left logical by variable amount
	dsrl reg,reg,imm	shift right logical
	dsra reg,reg,imm	shift right arithmetic
	dsrlv reg,reg,reg	shift right logical by variable amount
	dsrav reg,reg,reg	shift right arithmetic by variable amount
	转换
	cvt.d.l freg,freg	convert 64-bit integer to a double FP format
	cvt.l.d freg,freg	convert double FP to a 64-bit integer format
	标志
	slt reg,reg,reg	set if less than
	dslt reg,reg,reg	set if less than( 教材写错，上条为正确格式 )
	sltu reg,reg,reg	set if less than unsigned
	slti reg,reg,imm	set if less than immediate
	sltiu reg,reg,imm	set if less than immediate unsigned
	c.lt.d freg,freg	set FP flag if less than
	c.eq.d freg,freg	set FP flag if equal to
	c.le.d freg,freg	set FP flag if less than or equal to
	跳转
	j imm	jump to address                            (不会保存返回地址)
	jr reg	jump to address in register             （不会保存返回地址）
	jal imm	跳转到imm地址并且将返回地址记录在寄存器ra中（会保存）
	jalr reg	jump and link to address in register  （会保存）
	beqz reg,imm	branch if register is equal to zero
	bnez reg,imm	branch if register is not equal to zero
	beq reg,reg,imm	branch if pair of registers are equal
	bne reg,reg,imm	branch if pair of registers are not equal
	bc1t imm	branch to address if FP flag is TRUE
	bc1f imm	branch to address if FP flag is FALSE
	控制
	nop	no operation
	halt	stops the program