Re: Question on performance
On Oct 15 2010, Vincent Lefevre wrote:
What I meant is that isnan() and x != x can be seen as equivalent
even if __STDC_IEC_559__ is 1.
Don't you mean 'is not'? Anyway, my point is that it is not so, though
you could perfectly reasonably argue that the intention was that it
should be, and the discrepancies are defects in C99.
AFAIK, the C standard doesn't
provide a specified way to create a signaling NaN.
I wasn't talking about them. You are correct. My statement stands
with regard to quiet NaNs.
Regards,
Nick Maclaren.