Developer Guide and Reference

Contents

mlong-double

Lets you override the default configuration of the long double data type.

Syntax

Linux:
-mlong-double-
n
macOS:
None
Windows:
None
Arguments
n
Specifies the size of the long double data type. Possible values are:
64
Specifies that the size of the long double data type is 64 bits.
80
Specifies that the size of the long double data type is 80 bits. This is the default.
128
Specifies that the size of the long double data type is 128 bits.
Default
-mlong-double-80
Specifies that the size of the long double data type is 80 bits.
Description
This option lets you override the default configuration of the long double data type.
When you specify
-mlong-double-64
, the size of the long double data type is 8 bytes and the macro __LONG_DOUBLE_64__ is defined.
When you specify
-mlong-double-80
, the size of the long double data type is 12 bytes on IA-32 architecture and 16 bytes on Intel® 64 architecture.
This option has no effect on floating-point significand precision. That must be specified by using the
-pc64
or
-pc80
option.
Note that this option has no effect when you pass arguments. When you pass arguments, the 64-bit long double data type is treated as the double data type and it is always 64-bit.
Remember to include the math.h and complex.h header files when you use this option.
The following restrictions apply to this option:
  • __bultin_* functions using the long double type should not be used in the non-default mode with Intel compiler libraries.