X86 Architecture Overview

IA-32

and x86-64 The

two massively popular architectures IA-32 and x86-84 are so common that 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 Guide. The following diagram is taken directly from Chapter 3 of this volume:

Records Application

programmers generally use only general-purpose registers, floating-point registers, XMM registers, and YMM

registers.

General

purpose registers These are 64 bits wide and are used for integer arithmetic and logic, and to keep data and pointers in memory. The records are called R0… R15. Also:

You can access the lower 32-bit order of each register using the R0D names… R15D. The “D” stands for “double word” because strangely, the word “word” on this platform refers to a number of 16 bits. Why? Backward compatibility! The x86-64 grew out of a family of 16-bit processors created in the 1970s.
You can access the 16 lower-order bits of each register using the names R0W… R15W.
You can access the 8 lower-order bits of each register using the names R0B… R15B.
A0… R7 have aliases RAX, RCX, RBX, RDX, RSP, RBP, RSI, RDI, respectively.
R0D… R7D has 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 has aliases AL, CL, BL, DL, SPL, BPL, SIL, DIL, respectively.

RIP and RFLAGS RIP

the instruction pointer and RFAGS is the flag 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 called XMM0… XMM15. Use them for floating-point arithmetic and integers. You can perform operations on 128-bit integers, but you can also take advantage of its ability to perform

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

Parallel YMM registers

These are 256 bits wide. They are called YMM0… YMM15. Use them for floating-point arithmetic. It can do:

Four 64-bit floating-point operations in parallel Eight 32-bit floating-point operations in parallel

and a few other crazy things

FPU Registers

There are eight registers used to calculate with 80-bit

floating-point

values. Records do not have names because they are used in stack form.

Other

registers Application programmers can remain oblivious to the rest of the

registers:

The 8 control registers of the 32-bit processor: CR0, CR1, CR2, CR3, CR4, CR5, CR6, CR7. The bottom 16 bits of CR0 are called machine status word (MSW).

The

4 16-bit tables record: GDTR, IDTR, LDTR and TR

. The 8 32-bit debug registers: DR0, DR1, DR2, DR3, DR4, DR5, DR6 and DR7. The

5 test records: TR3, TR4, TR5, TR6 and TR7. The
range of memory types

records machine-specific

registers

Machine check records instruction set See Volume 1

SDM Chapter 5 for a good overview of

all processor instructions and Volume 2 for complete information

. The following table lists most of the

available statements, using the instruction names specified in Intel syntax. Not all processors support all instructions, of course.

The vertical bar means OR, the brackets mean OPTIONAL, and the 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.

Memory addressing

In protected mode, applications can choose a flat or segmented memory model (see SDM Volume 1, Chapter 3 for more information); only a 16-bit segmented model is available in real mode. 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 six segment registers; The base is one of 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 minimum reference consists of only a base record or only one offset; A scale can only appear if an index is present.

Sometimes the memory reference is written like this:

Selector Scroll (Base, Index,

Scale) Data Types

Data Types are

Type NameNumber of bitsBitIndexes Byte87.. 0 Word1615.. 0 Doubleword3232.. 0 Quadword6463.. 0 Doublequadword128127.. 0

Small endianness

The IA-32 is little endian, which means that 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 in the part lower, Then it is easier to read these values!

Flag registration

Many instructions cause the flag register to be updated. For example, if you execute an add statement and the sum is too large to fit in the target record, the Overflow flag is set.

3 3 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 0 0 0 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 indicators are described in Section 3.4.3 of Volume 1 of the SDM. To determine how each instruction affects indicators, refer to SDM Volume 1, Appendix A.

The

System Developer’s Manual contains a wealth of important information and is required reading for all assembly language programmers and backend compiler writers. The manual is divided into several volumes; Links to all volumes are here. Highlights

of Volumes 1 and 2:

Volume 1: Chapter
- 1 – About this manual
- IA-32 and Intel 64 architectures, a description of many of the microarchitectures, processors, and technologies
- 3 – Basic Runtime Environment
- 4 – Data Types Chapter
- 5 – Summary of the instruction set. List all instructions and a brief (but not precise) description of each. The instructions are grouped into convenient categories.
- 6 – Details on Calls and Returns
- SSSE3, SSE4 and AESNI Instructions Chapter 13 – XSAVE Chapter 14
- –
- , FMA Instructions and AVX2
- 15 – AVX-512
- Transactional Synchronization
- 18 – All About I/O Instructions Chapter 19
- How to Determine Which Processor You Have and What Its Features
- A – Shows which instructions affect which indicators
- B – Condition Codes
- Floating-Point
- Exceptions
- E – Guidelines for Writing SIMD Exception Handlers

Volume 2: Chapter 1 – About

- this manual
- Instruction Formats
- Chapter 3-5 – Instruction Set Reference: Complete Description and Encodings of Each Statement with Names Beginning with A — L Chapter
- 6 – Safer Mode Extensions Reference
- A – Opcode Map
- B – Encoding Summary Appendix
- C – Compiler Intrinsics

Blogs