[cfp-interest 3748] Re: INFINITY on a system supporting _Float16 - not urgent

[Damian McGuckin]

On Mon, 22 Dec 2025, Damian McGuckin wrote:

> What is the semantics of a comparison of INFINITY against a _Float16?

Previous, INFINITY has a type (float) which is the smallest supported 
floating type so that it was only ever promoted in type if there was a 
need for conversion.

Does that mean that its type needs to change when _FLoat16 is supported?
What about is the proposed IEEE P3109 8-bit float ever permeates into C
and it has a _Float8 type?

Thanks - Damian