The ARC cpu family currently does not have hardware floating point support. Software floating point support is provided by GCC and uses IEEE floating-point numbers.
GCC
Go to the first, previous, next, last section, table of contents.