This option controls generating floating-point calculation instructions.
[Specification format]
- | Interpretation when omitted |
If the-Xcpu=g3k option is specified, -Xfloat=soft is assumed.
In any other case, -Xfloat=fpu is assumed.
[Detailed description]
- | This option controls generating floating-point calculation instructions. |
- | The items that can be specified as type are shown below.
An error will occur if any other item is specified. |
|
soft
|
Generates runtime function call instructions for floating-point calculations.
|
|
fpu
|
Generates floating-point calculation instructions of FPU (floating-point unit) for floating-point calculations.
However, if -Xcpu=g3kh is specified, runtime function call instructions for double-precision calculations will be generated.
If this option is specified together with the -Xcpu=g3k option, this option is invalid and -Xfloat=soft is assumed.
|
- | An error will occur if type is omitted. |
- | If this option is specified together with the -Xcpu=g3k option, this option will be invalid and runtime function call instructions will always be generated. |
- | If soft is specified as type, the -Xround=zero option will be invalid and -Xround=nearest will always be valid. |
[Example of use]
- | To generate the jarl32 and jr32 instructions for function-call branches, describe as: |
>ccrh -Xfloat=soft -Xcommon=rh850 main.c
|