[Cfp-interest] Fw: 18661-2 and 754

[Jim Thomas]

For these operations 754 says "they only affect the sign bit”. Does this make sense for extended formats, given that 754 doesn’t specify their encodings?

> On May 15, 2018, at 3:24 PM, David Hough CFP <pcfp at oakapple.net> wrote:
> 
>> Not all IEEE 754 operations guarantee a canonical result, do they?  Eg,=0D
> copying, negating, absolute value. 
> 
> The result is as canonical or non-canonical 
> as the input, differing only in the sign bit.