[compiler-rt] HP FP support

Project Infrastructure LLVM Dev List Archives
1

Is there any reason why the HP FP conversion and comparison builtins are not defined?

Thank you,

2

Hi Evandro

Is there any reason why the HP FP conversion and comparison builtins are
not defined?

As far as I can tell, this hasn't been a requirement for many targets,
but is becoming more of a requirement recently. clang itself doesn't
support _Float16 or __fp16 for all targets, so even if compiler-rt
support is available, the predominant llvm frontend does yet produce
them for most targets.

I recently floated [a patch](⚙ D84877 Support for soft fp16 to fp64 IEEE conversions) to add
support for the `__extendhfdf2` (promote half to double) builtin. Your
feedback would be welcome.

The infrastructure is mostly there already, so I'd encourage you to add
more features as needed.

All the Best

Luke