This option determines the precision of floating-point results and the location at which floating-point interrupts are reported.
.-IMPRECISE---. >>-+-NOIMPRECISE-+---------------------------------------------><
ABBREVIATIONS: IMP, NIMP
Use of this option produces smaller object code that runs faster. It is recommended for your production programs.
Although NOIMPRECISE does provide better floating-point error detection than IMPRECISE, the Windows operating system does not allow immediate detection of floating-point exceptions. If you have a statement in your program that is likely to raise a floating-point exception, you can avoid this detection problem by enclosing the statement, by itself, in a BEGIN block.