On Sun, Aug 14, 2016 at 12:05:37PM +0000, Mark Dickinson wrote:
> But I don't think there's a real problem here so long as you don't 
> have an expectation of getting super-accurate (e.g., correctly rounded 
> or faithfully rounded) results; testing that results are accurate to 
> within 10 ulps or so is probably good enough.

Thanks for the advice Mark.

This has certainly showed my naivety when it comes to floating point :-( 
I thought IEEE 754 was supposed to put an end to these sorts of cross- 
platform differences. If this was a trig function, I wouldn't have been 
so surprised, but I was not expecting pow() to differ like this. Once I 
started getting exact results on my machine, I thought everyone would 
see the same thing.