background image

A-6 Vol. 2D

OPCODE MAP

A.2.5  

Superscripts Utilized in Opcode Tables

Table A-1 contains notes on particular encodings. These notes are indicated in the following opcode maps by super-
scripts. Gray cells indicate instruction groupings.

A.3 

ONE, TWO, AND THREE-BYTE OPCODE MAPS

See Table A-2 through Table A-5 below. The tables are multiple page presentations. Rows and columns with 
sequential relationships are placed on facing pages to make look-up tasks easier. Note that table footnotes are not 
presented on each page. Table footnotes for each table are presented on the last page of the table.

Table A-1.  Superscripts Utilized in Opcode Tables

Superscript

Symbol

Meaning of Symbol

1A

Bits 5, 4, and 3 of ModR/M byte used as an opcode extension (refer to Section A.4, “Opcode Extensions For One-Byte 

And Two-byte Opcodes”).

1B

Use the 0F0B opcode (UD2 instruction) or the 0FB9H opcode when deliberately trying to generate an invalid opcode 

exception (#UD).

1C

Some instructions use the same two-byte opcode. If the instruction has variations, or the opcode represents 

different instructions, the ModR/M byte will be used to differentiate the instruction. For the value of the ModR/M 

byte needed to decode the instruction, see Table A-6. 

i64

The instruction is invalid or not encodable in 64-bit mode. 40 through 4F (single-byte INC and DEC) are REX prefix 

combinations when in 64-bit mode (use FE/FF Grp 4 and 5 for INC and DEC).

o64

Instruction is only available when in 64-bit mode.

d64

When in 64-bit mode, instruction defaults to 64-bit operand size and cannot encode 32-bit operand size. 

f64

The operand size is forced to a 64-bit operand size when in 64-bit mode (prefixes that change operand size are 

ignored for this instruction in 64-bit mode).

v

VEX form only exists. There is no legacy SSE form of the instruction. For Integer GPR instructions it means VEX 

prefix required.

v1

VEX128 & SSE forms only exist (no VEX256), when can’t be inferred from the data size.