Intel 253666-024US Manuel D’Utilisation

Page de 760
3-346 Vol. 2A
FLDCW—Load x87 FPU Control Word
INSTRUCTION SET REFERENCE, A-M
FLDCW—Load x87 FPU Control Word
Description
Loads the 16-bit source operand into the FPU control word. The source operand is a 
memory location. This instruction is typically used to establish or change the FPU’s 
mode of operation.
If one or more exception flags are set in the FPU status word prior to loading a new 
FPU control word and the new control word unmasks one or more of those excep-
tions, a floating-point exception will be generated upon execution of the next 
floating-point instruction (except for the no-wait floating-point instructions, see the 
section titled “Software Exception Handling” in Chapter 8 of the Intel® 64 and IA-32 
Architectures Software Developer’s Manual, Volume 1
). To avoid raising exceptions 
when changing FPU operating modes, clear any pending exceptions (using the FCLEX 
or FNCLEX instruction) before loading the new control word.
This instruction’s operation is the same in non-64-bit modes and 64-bit mode.
Operation
FPUControlWord ← SRC;
FPU Flags Affected
C0, C1, C2, C3 undefined.
Floating-Point Exceptions
None; however, this operation might unmask a pending exception in the FPU status 
word. That exception is then generated upon execution of the next “waiting” floating-
point instruction.
Protected Mode Exceptions
#GP(0)
If a memory operand effective address is outside the CS, DS, 
ES, FS, or GS segment limit.
If the DS, ES, FS, or GS register is used to access memory and it 
contains a NULL segment selector.
#SS(0)
If a memory operand effective address is outside the SS 
segment limit.
#NM
CR0.EM[bit 2] or CR0.TS[bit 3] = 1.
#PF(fault-code)
If a page fault occurs.
Opcode
Instruction
64-Bit 
Mode
Compat/
Leg Mode
Description
D9 /5
FLDCW m2byte
Valid
Valid
Load FPU control word from m2byte.