TestFDIV Variable
The TestFDIV
variable tells whether the
floating-point processor has a correct implementation of the FDIV
instruction.
If you use the $U
or
$SafeDivide
compiler directive, all floating-point
division operations call a function that tests the floating-point
processor for the infamous Pentium FDIV bug. If the bug is present,
future divisions will be performed by a software subroutine that
gives the correct answer. If the bug is not present, future divisions
will be performed by the floating-point hardware.
The TestFDIV
variable contains one of the
following values:
The processor is flawed and will be bypassed.
The processor has not yet been tested.
The processor has been tested and is correct.