Make C floating point literals float (rather than double)

-fsingle-precision-constant flag can be used. It causes floating-point constants to be loaded in single precision even when this is not exact.

Note- This will also use single precision constants in operations on double precision variables.