AMD Typewriter x86 사용자 설명서

22007E/0—November 1999

This chapter describes the MacroOPs generated by each
decoded instruction, along with the relative static execution
latencies of these groups of operations. Tables 19 through 24
starting on page 188 define the integer, MMX™, MMX
extensions, floating-point, 3DNow!™, and 3DNow! extensions
instructions, respectively. 

The first column in these tables indicates the instruction
mnemonic and operand types with the following notations:

■

reg8—byte integer register defined by instruction byte(s) or 
bits 5, 4, and 3 of the modR/M byte

■

mreg8—byte integer register defined by bits 2, 1, and 0 of 
the modR/M byte

■

reg16/32—word and doubleword integer register defined by 
instruction byte(s) or bits 5, 4, and 3 of the modR/M byte

■

mreg16/32—word and doubleword integer register defined 
by bits 2, 1, and 0 of the modR/M byte

■

mem8—byte memory location

■

mem16/32—word or doubleword memory location

■

mem32/48—doubleword or 6-byte memory location

■

mem48—48-bit integer value in memory

■

mem64—64-bit value in memory

■

imm8/16/32—8-bit, 16-bit or 32-bit immediate value

■

disp8—8-bit displacement value