Support of the IEEE 754-2008 32/64/128 bit floating point types.
Support of the 4 elementary operations plus square root, fused
multiply-add, minimum, maximum, conversion to and from 32/64/128
bit signed or unsigned integers.
RISC-V FPU specifics are supported by default.
128 bit floating point type and 128 bit integers relying on the
C compiler __int128 type.