https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121525
--- Comment #1 from Segher Boessenkool <segher at gcc dot gnu.org> --- (In reply to Michael Meissner from comment #0) > Power9 (i.e. ISA 3.0) added support for the xscvdphp, xscvhpdp instructions > that convert scalar IEEE 16-bit floating points to/from SFmode as well as > the xvcvsphp, and xvcvhpsp instructions that convert vector 16-bit floats > to/from V4SF mode. > > GCC 16 should add support for the _Float16 type and generate these > instructions. (Those instructions convert between HP and DP). Yes, there should be a generic type for IEEE HP float, that we can then support. This should be added to the ABIs if you want to use it as function arguments.
