Comment Re:Fixed-point arithmetic (Score 1) 226
In IEEE floating point math, "(a+b)+c" might not be the same as "a+(b+c)".
And if your code is sensitive to that then you are using the wrong algorithm. So it's the algorithm which needs to be fixed. not the FPU environment.
All floating[-point work is approximate. It's up to you to ensure that the significances are greater than the approximations.