This option changes the interpretation of the double and long double types.
Both the double type and long double type are handled as the single-precision floating-point type (as when the -dbl_size=4 option is specified).
When 4 is specified for the parameter, the double type and the long double type are handled as the single-precision floating-point type. |
When 8 is specified for the parameter, the double type and the long double type are handled as the double-precision floating-point type. |
When the -dbl_size=4 option is specified and a standard library function for the double type is called, it is replaced with the corresponding standard library for the float type. |