Add minNum and maxNum from IEEE 754-2008 (pull request, with patch)
I was unable to open a pull request on this odd Gitlab of yours, so here is a patch! minmax.patch
It adds minimum and maximum operations to IEEE754/Binary.v.
The operations are modeled after the minNum and maxNum operations of IEEE 754-2008.
In particular, if one of the arguments is NaN but not both, the non-NaN argument is returned as result.
If one argument is +0.0 and the other -0.0, the standard does not specify which zero is returned. We return -0.0 as minimum and +0.0 as maximum, consistently with several hardware implementations (AArch64, RISC-V).