For floating-point numbers the C/C++ compiler generates code in /390 format or IEEE format, as required. You specify the format you want by means of the FP-ARITHMETICS clause of the MODIFY-MODULE-PROPERTIES compiler option.
MODIFY-MODULE-PROPERTIES -
...
FP-ARITHMETICS= { *390-FORMAT | *IEEE-FORMAT }, -
LOWER-CASE-NAMES=*YES, -
SPECIAL-CHARACTERS=*KEEP, -
...
FP-ARITHMETICS=*390-FORMAT
The compiler generates code for constants and arithmetic operations in /390 format. *390-FORMAT is the default.
FP-ARITHMETICS=*IEEE-FORMAT
The compiler generates code for constants and arithmetic operations in IEEE format. In addition, the _IEEE preprocessor define is set to 1. Unless the _IEEE_SOURCE preprocessor define is set to 0 (see "Controlling the mapping of original functions to the associated IEEE variants"), the original /390 library functions are automatically mapped to the associated IEEE functions.
LOWER-CASE-NAMES=*YES
SPECIAL-CHARACTERS=*KEEP
By specifying these, you prevent:
the names of the IEEE functions (see "C library functions that support IEEE floating-point numbers") from being truncated to eight characters
lowercase letters from being converted to uppercase and the character “_” from being replaced by “$” in the function names
In POSIX you specify the IEEE format by means of the following option:
-K ieee_floats
To ensure the IEEE function names are processed correctly, you specify:
-K llm_keep
-K llm_case_lower