x86 Architecture Overview

6/28/22, 10:26 AM x86 Architecture Overview
x86 Architecture
Overview
The IA-32 is the instruction set architecture (ISA) of Intel’s most successful line of 32-bit processors, and the
Intel 64 ISA is its extension into 64-bit processors. (Actually Intel 64 was invented by AMD, who called it x86-64).
These notes summarize a few items of interest about these two ISAs. They in no way serve as a substitute for
reading Intel’s manuals.
CONTENTS
IA-32 and x86-64 • x86-64 Architecture Diagram • Registers • Instruction Set • Addressing Memory • Flags Register • The System Developer’s
Manual
IA-32 and x86-64

The two massively popular architectures IA-32 and x86-84 are so common, they are described in a single set of
manuals.
The following notes briefly summarize the latter architecture only.
x86-64 Architecture Diagram

The basic architecture of the x86-64 is described in Volume 1 of the System Developer’s Manual. The following
diagram is taken directly from Chapter 3 in this volume:
https://cs.lmu.edu/~ray/notes/x86overview/ 1/9
Registers
Application Programmers generally use only the general purpose registers, floating point registers, XMM, and YMM
registers.
General Purpose Registers

These are 64 bits wide and used for integer arithmetic and logic, and to hold both data and pointers to memory.
The registers are called R0...R15. Also:
You can access the lower order 32-bits of each register using the names R0D...R15D. The “D” stands for
“doubleword” because strangely, the word “word” on this platform refers to a 16-bit quantity. Why?
Backward compatibilty! The x86-64 grew out of a 16-bit processor family created in the 1970s.
You can access the lower order 16-bits of each register using the names R0W...R15W.
You can access the lower order 8-bits of each register using the names R0B...R15B.
R0...R7 have aliases RAX, RCX, RBX, RDX, RSP, RBP, RSI, RDI, respectively.
R0D...R7D have aliases EAX, ECX, EBX, EDX, ESP, EBP, ESI, EDI, respectively.
R0W...R7W have aliases AX, CX, BX, DX, SP, BP, SI, DI, respectively.
R0B...R7B have aliases AL, CL, BL, DL, SPL, BPL, SIL, DIL, respectively.
RIP and RFLAGS

RIP is the instruction pointer and RFAGS is the flags register.
Segment Registers
These are CS, DS, SS, ES, FS, and GS. I haven’t used them in 64-bit programming.
XMM Registers
These are 128-bits wide. They are named XMM0...XMM15. Use them for floating-point and integer arithmetic. You
can do operations on 128-bit integers, but you can also take advantage of their ability to do operations in parallel:
Two 64-bit integer operations in parallel

Four 32-bit integer operations in parallel
Eight 16-bit integer operations in parallel
Sixteen 8-bit integer operations in parallel
Two 64-bit floating-point operations in parallel
Four 32-bit floating-point operations in parallel
YMM Registers
These are 256-bits wide. They are named YMM0...YMM15. Use them for floating-point arithmetic. You can do:
Four 64-bit floating-point operations in parallel

Eight 32-bit floating-point operations in parallel
and some other crazy things.
FPU Registers
There are eight registers used for computing with 80-bit floating point values. The registers don’t have names
because they are used in a stack-like fashion.
Other Registers
Application programmers can remain oblivious of the rest of the registers:
The 8 32-bit processor control registers: CR0, CR1, CR2, CR3, CR4, CR5, CR6, CR7. The lower 16 bits of CR0 is
called the Machine Status Word (MSW).
The 4 16-bit table registers: GDTR, IDTR, LDTR and TR.
The 8 32-bit debug registers: DR0, DR1, DR2, DR3, DR4, DR5, DR6 and DR7.
The 5 test registers: TR3, TR4, TR5, TR6 and TR7.
The memory type range registers
The machine specific registers
The machine check registers
Instruction Set
See the SDM Volume 1, Chapter 5 for a nice overview of all of the processor instructions and Volume 2 for
complete information.
The following table shows most of the available instructions, using the instruction names as specified in the Intel
syntax. Not every processor supports every instruction, of course.
The vertical bar means OR, the square brackets mean OPTIONAL, and parentheses are used for grouping. For
example:
SH(L|R)[D] stands for SHL , SHR , SHLD , SHRD .

PUSH[A[D]] stands for PUSH , PUSHA , PUSHAD .
INTEGER FPU SSE SSE2

MOV F[I]LD MOV(A|U)PS MOV(A|U)PD
CMOV[N]((L|G|A|B)[E]|E|Z|S|C|O|P) F[I]ST[P] MOV(H|HL|L|LH)PS MOV(H|L)PD
XCHG FBLD MOVSS MOVSD
BSWAP FBSTP MOVMSKPS MOVMSKPD
XADD FXCH
CMPXCHG[8B] FCMOV[N](E|B|BE|U) ADD(P|S)S ADD(P|S)D
PUSH[A[D]] | POP[A[D]] SUB(P|S)S SUB(P|S)D
IN | OUT FADD[P] MUL(P|S)S MUL(P|S)D
CBW | CWDE | CWD | CDQ FIADD DIV(P|S)S DIV(P|S)D
MOVSX | MOVZX FSUB[R][P] RCP(P|S)S SQRT(P|S)D
FISUB[R] SQRT(P|S)S MAX(P|S)D
ADD | ADC FMUL[P] RSQRT(P|S)S MIN(P|S)D
SUB | SBB FIMUL MAX(P|S)S
[I]MUL FDIV[R][P] MIN(P|S)S CMP(P|S)D
[I]DIV FIDIV[R] [U]COMISD
INC | DEC FPREM[1] CMP(P|S)S
NEG FABS [U]COMISS ANDPD
CMP FCHS ANDNPD
FRNDINT ANDPS ORPD
DAA | DAS FSCALE ANDNPS XORPD
AAA | AAS | AAM | AAD FSQRT ORPS
FXTRACT XORPS SHUFPD
AND | OR | XOR | NOT UNPCK(H|L)PD
F[U]COM[P][P] SHUFPS
SH(L|R)[D] FICOM[P] UNPCK(H|L)PS CVT(PI|DQ)2PD
SA(L|R) F[U]COMI[P] CVT[T]PD2(PI|DQ)
RO(L|R) FTST CVTPI2PS CVTSI2SD
RC(L|R) FXAM CVT[T]PS2PI CVT[T]SD2SI
CVTSI2SS CVTPS2PD
BT[S|R|C] FSIN CVT[T]SS2SI CVTPD2PS
BS(F|R) FCOS CVTDQ2PS
SET[N]((L|G|A|B)[E]|E|Z|S|C|O|P) FSINCOS PAVG(B|W) CVT[T]PS2DQ
TEST FPTAN PEXTRW CVTSS2SD
FPATAN PINSRW CVTSD2SS
JMP F2XM1 P(MIN|MAX)(UB|SW)
J[N]((L|G|A|B)[E]|E|Z|S|C|O|P) FYL2X PMOVMSKB MOVDQ(A|U)
J[E]CXZ FYL2XP1 PMULHUW MOVQ2DQ
LOOP[N][Z|E] PSADBW MOVDQ2Q
CALL | RET FLD1 PSHUFW PUNPCK(H|L)QDQ
INT[O] | IRET FLDZ PADDQ
ENTER | LEAVE FLDPI LDMXCSR PSUBQ
BOUND FLDL2E STMXCSR PMULUDQ
FLDLN2 PSHUF(LW|HW|D)
MOVS[B|W|D] FLDL2T MASKMOVQ PS(L|R)LDQ
CMPS[B|W|D] FLDLG2 MOVNT(Q|PS)
SCAS[B|W|D] PREFETCHT(0|1|2) MASKMOVDQU
LODS[B|W|D] FINCSTP PREFETCHNTA MOVNT(PD|DQ|I)
STOS[B|W|D] FDECSTP SFENCE CLFLUSH
INS[B|W|D] FFREE LFENCE
OUTS[B|W|D] F[N]INIT MFENCE
REP[N][Z|E] F[N]CLEX PAUSE
F[N]STCW
STC | CLC | CMC FLDCW
STD | CLD F[N]STENV
STI | CLI FLDENV
LAHF | SAHF F[N]SAVE
PUSHF[D] | POPF[D] FRSTOR
F[N]STSW
LDS | LES | LFS | LGS | LSS FWAIT | WAIT
FNOP
LEA
NOP FXSAVE
UD2 FXRSTOR
XLAT[B]
CPUID
SYSTEM MMX SSE3 SSE4

LGDT | SGDT MOVD FISTTP PMUL(LD|DQ)
LLDT | SLDT MOVQ DPP(D|S)
LTR | STR LDDQU
LIDT | SIDT PACKSS(WB|DW) MOVNTDQA
LMSW | SMSW PACKUSWB ADDSUBP(S|D)
CLTS PUNPCK(H|L)(BW|WD|DQ) BLEND[V](PD|PS)
ARPL HADDP(S|D) PBLEND(VB|W)
LAR PADD(B|W|D) HSUBP(S|D)
LSL PADD(S|US)(B|W) PMIN(UW|UD|SB|SD)
VERR | VERW PSUB(B|W|D) MOVS(H|L)DUP PMAX(UW|UD|SB|SD)
INVD | WBINVD PSUB(S|US)(B|W) MOVDDUP
INVLPG PMUL(H|L)W ROUND(P|S)(S|D)
LOCK PMADDWD MONITOR
HLT PCMP(EQ|GT)(B|W|D) MWAIT EXTRACTPS
RSM INSERTPS
RDMSR | WRMSR PAND PINSR(B|D|Q)
RDPMC PANDN PEXTR(B|W|D|Q)
RDTSC POR PMOV(S|Z)X(BW|BD|WD|BQ|WQ|DQ)
SYSENTER PXOR
SYSEXIT MPSADBW
PS(L|R)L(W|D|Q)
PSRA(W|D) PHMINPOSUW
EMMS PTEST
PCMPEQQ
PACKUSDW
PCMP(E|I)STR(I|M)
PCMPGTQ
CRC32
POPCNT
64-BIT MODE VIRTUAL MACHINE SSSE3 AESNI

CDQE VMPTRLD PHADD(W|SW|D) AESDEC[LAST]
CMPSQ VPTRST PHSUB(W|SW|D) AESENC[LAST]
CMPXCHG16B VMCLEAR PABS(B|W|D) AESIMC
LODSQ VMREAD PMADDUBSW AESKEYGENASSIST
MOVSQ VMWRITE PMULHRSW PCLMULQDQ
MOVZX VMCALL PSHUFB
STOSQ VMLAUNCH PSIGN(B|W|D)
SWAPGS VMRESUME PALIGNR
SYSCALL VMXOFF
SYSRET VMXON
INVEPT
INVVPID
Addressing Memory
In protected mode, applications can choose a flat or segmented memory model (see the SDM Volume 1, Chapter
3 for details); in real mode only a 16-bit segmented model is available. Most programmers will only use protected
mode and a flat-memory model, so that’s all we’ll discuss here.
A memory reference has four parts and is often written as
[SELECTOR : BASE + INDEX * SCALE + OFFSET]
The selector is one of the six segment registers; the base is one of the eight general purpose registers; the index is
any of the general purpose registers except ESP; the scale is 1, 2, 4, or 8; and the offset is any 32-bit number.
(Example: [fs:ecx+esi*8+93221] .) The minimal reference consists of only a base register or only an offset; a
scale can only appear if there is an index present.
Sometimes the memory reference is written like this:
selector
offset(base,index,scale)
Data Types
The data types are
Type name Number of bits Bit indices
Byte 8 7..0
Word 16 15..0
Doubleword 32 32..0
Quadword 64 63..0
Doublequadword 128 127..0
Little Endianness
The IA-32 is little endian, meaning the least significant bytes come first in memory. For example:
0 12
1 31 byte @ 9 = 1F
2 CB word @ B = FE06
3 74 word @ 6 = 230B
4 67 word @ 1 = CB31
5 45 dword @ A = 7AFE0636
6 0B qword @ 6 = 7AFE06361FA4230B
7 23 word @ 2 = 74CB
8 A4 qword @ 3 = 361FA4230B456774
9 1F dword @ 9 = FE06361F
A 36
B 06
C FE
D 7A
E 12
Note that if you draw memory with the lowest bytes at the bottom, then it is easier to read these values!
Flags Register
Many instructions cause the flags register to be updated. For example if you execute an add instruction and the
sum is too big to fit into the destination register, the Overflow flag is set.
3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0
1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
+---------------------------------------------------------------+
| | | | | | | | | | |I|V|V|A|V|R| |N| I |O|D|I|T|S|Z| |A| |P| |C|
| | | | | | | | | | |D|I|I|C|M|F| |T| P |F|F|F|F|F|F| |F| |F| |F|
| | | | | | | | | | | |P|F| | | | | | L | | | | | | | | | | | | |
+---------------------------------------------------------------+
The flags are described in Section 3.4.3 of Volume 1 of the SDM. To determine how each instruction affects the
flags, see Appendix A of Volume 1 of the SDM.
The System Developer’s Manual

The System Developer’s Manual contains vast amounts of important information and is required reading for all
assembly language programmers and backend compiler writers. The manual is split into several volumes; links to
all volumes are here. Highlights from Volumes 1 and 2:
Volume 1:
Chapter 1 - About this manual
Chapter 2 - History of the IA-32 and Intel 64 architectures, a description of many of the
microarchitectures and processors and technologies
Chapter 3 - Basic execution environment
Chapter 4 - Data types
Chapter 5 - Instruction set summary. Lists all instructions and a brief (but not precise) description of
each. Instructions are grouped into convenient categories.
Chapter 6 - Details on calls and returns, and exceptions
Chapter 7 - All about general purpose instructions
Chapter 8 - All about FPU instructions
Chapter 9 - All about MMX instructions
Chapter 10 - All about SSE instructions
Chapter 11 - All about SSE2 instructions
Chapter 12 - All about SSE3, SSSE3, SSE4 and AESNI instructions
Chapter 13 - XSAVE
Chapter 14 - All about AVX, FMA, and AVX2 instructions
Chapter 15 - AVX-512
Chapter 16 - All about transactional synchronization instructions
Chapter 17 - Memory protection extensions
Chapter 18 - All about I/O instructions
Chapter 19 - How to determine what processor you have, and what its features are
Appendix A - Shows which instructions affect which flags

Appendix B - Condition codes
Appendix C - Floating-point exceptions
Appendix D - Guidelines for writing x87 exception handlers
Appendix E - Guidelines for writing SIMD exception handlers
Volume 2:
Chapter 1 - About this manual
Chapter 2 - Instruction formats
Chapter 3-5 - Instruction set reference: full description, and encodings, of every instruction with
names beginning with A — L
Chapter 6 - Safer Mode Extensions Reference
Appendix A - Opcode map
Appendix B - Encoding summary
Appendix C - Compiler intrinsics

x86 Architecture Overview

Uploaded by

x86 Architecture Overview

Uploaded by

6/28/22, 10:26 AM x86 Architecture Overview

IA-32 and x86-64

The following notes briefly summarize the latter architecture only.

x86-64 Architecture Diagram

General Purpose Registers

RIP and RFLAGS

Two 64-bit integer operations in parallel

Four 32-bit floating-point operations in parallel

Four 64-bit floating-point operations in parallel

and some other crazy things.

SH(L|R)[D] stands for SHL , SHR , SHLD , SHRD .

INTEGER FPU SSE SSE2

SYSTEM MMX SSE3 SSE4

64-BIT MODE VIRTUAL MACHINE SSSE3 AESNI

A memory reference has four parts and is often written as

[SELECTOR : BASE + INDEX * SCALE + OFFSET]

Sometimes the memory reference is written like this:

Type name Number of bits Bit indices

The System Developer’s Manual

Appendix A - Shows which instructions affect which flags

You might also like