| Syntax: | Default: |
| gen(longDouble[, yes]) gen(longDouble, no) |
This option controls whether long doubles are represented as 128-bit values or 64-bit values.
By default, gen(longDouble, no) is set, and long doubles are represented by 64-bit values.
Separate libraries are provided that support 128-bit long double types. These libraries will be automatically linked if you specify the defaults(string) option using any string with the 128 suffix, such as xlC128, xlc128, or cc128. You can also explicitly link to the 128-bit versions of the libraries by specifying them as sources in your configuration file. The 128-bit versions of the libraries that correspond to the 64-bit libraries are shown in the following table:
| Default (64-bit) long double library | 128-bit long double library |
|---|---|
| libC.a | libC128.a |
| libCns.a | libC128ns.a |
| libC_r.a | libC128_r.a |
| libCns_r.a | libC128ns_r.a |
| libc.a | libc128.a |
| libc_r.a | libc128_r.a |
Linking without the 128-bit versions of the libraries when your program uses 128-bit long doubles, for example, if you specify gen(longDouble) alone, may produce unpredictable results.
The gen(longDouble) option defines __LONGDOUBLE128.
This is a global codestore option.
![]()
Incremental C++ Build
C Compilation
Configuration Files
![]()
Set Build Options for
the Project
Build
![]()
Code Generation Options
![]()
Code Generation Options
