nextafter

[David G. Hough on validgh]

> it's not clear nextafter deserves any architectural support
> at all, given how rarely it is used

Of course it's rarely used because it has no support in the hardware or
programming language architectures, or the programming paradigms for
environmental inquiries.  Of course to meet RISC criteria you'd have to
show at least that it would or could be used in inner loops of some sort
for which efficiency made a measurable difference.