x86-64 Cheat Sheet

General-Purpose Registers

Name	64-bit	32-bit	16-bit	8-bit	Preserved	Usage
Accumulator	rax	eax	ax	ah:al		return
Base	rbx	ebx	bx	bh:bl	✓
Counter	rcx	ecx	cx	ch:cl		arg4
Data	rdx	edx	dx	dh:dl		arg3
Source	rsi	esi	si	sil		arg2
Destination	rdi	edi	di	dil		arg1
	r8	r8d	r8w	r8b		arg5
	r9	r9d	r9w	r9b		arg6
	r10	r10d	r10w	r10b
	r11	r11d	r11w	r11b
	r12	r12d	r12w	r12b	✓
	r13	r13d	r13w	r13b	✓
	r14	r14d	r14w	r14b	✓
	r15	r15d	r15w	r15b	✓
Base Pointer	rbp	ebp	bp	bpl	✓	function stack base (optional)
Stack Pointer	rsp	esp	sp	spl	✓	top of stack *
Instruction Pointer	rip	eip	ip

* The rsp register must be 16-byte aligned (i.e. divisible by 16) before making a call.

Label	Name	Description
CF	carry flag	last arithmetic operation carried/borrowed
ZF	zero flag	result of last operation was zero
SF	sign flag	result of last operation was negative
OF	overflow flag	last result had a signed overflow

Instruction	Test	After `TEST x, x`	After `CMP y, x`
je	ZF	x = 0	x = y
jne	~ZF	x ≠ 0	x ≠ y
ja	~CF & ~ZF		x > y (unsigned)
jae	~CF		x ≥ y (unsigned)
jb	CF		x < y (unsigned)
jbe	CF \| ZF		x ≤ y (unsigned)
jg	~ZF & SF=OF	x > 0 (signed)	x > y (signed)
jge	SF=OF	x ≥ 0 (signed)	x ≥ y (signed)
jl	SF≠OF	x < 0 (signed)	x < y (signed)
jle	ZF \| SF≠OF	x ≤ 0 (signed)	x ≤ y (signed)
jc	CF
jnc	~CF
jo	OF
jno	~OF
js	SF
jns	~SF

Floating point registers st(0) to st(7) are manipulated by the x87 instructions, which we won't be using.

We will be using these registers and the instructions that work on them for floating point operations:

512-bit	256-bit	128-bit	Usage
zmm0	ymm0	xmm0	arg1, return
zmm1	ymm1	xmm1	arg2
⋮			⋮
zmm7	ymm7	xmm7	arg8
zmm8	ymm8	xmm8
⋮
zmm15	ymm15	xmm15

Memory access in an instruction: displacement(base, index, scale)

e.g. -16(%rbp, %rdx, 8) == rbp + (rdx * 8) - 16. In C, that might be rbp[rdx-2] if you're working with elements of 8 bytes.

Name	Size	Suffix	Data
byte	8-bits	`b`	`.byte`
word	16-bits	`w`	`.word`
long word	32-bits	`l`	`.long`
quad word	64-bits	`q`	`.quad`

Integer movement: mov, push, pop, movabs, movsx, movzx.
Unsigned integer arithmetic: add, sub, mul, div, shl, shr.
Signed integer arithmetic: add, sub, imul, idiv, sal, sar, neg.
Bit manipulation: and, or, not.
Integer comparison: cmp, test
Jumps: jmp, jX, jnX, call, ret.
Pointer creation: lea.
Conditional move: cmovX.
Double-precision scalars: movq, addsd, subsd, mulsd, divsd, comisd, pxor.
Single-precision scalars: movd, addss, subss, mulss, divss, comiss, pxor.
Double-precision vectors: vmovupd, vaddpd, vsubpd, vmulpd, vdivpd, vxorpd, vbroadcastsd.
Single-precision vectors: vmovups, vaddps, vsubps, vmulps, vdivps, vxorps, vbroadcastss.