On 04/02/2012 09:11 AM, René Nyffenegger wrote:
I hope this is the right place to ask this question. I only recently tried to compile gcc and it is obviously dependant on mpfr, mpc and gmp. These seem to be high precision mathematical libraries. Now, I am wondering why a compiler not only needs one of these but even three of them.
While optimizing code, GCC can simplify some floating point expressions. It uses these libraries to carry out the arithmetic associated with these optimizations.
For example, things like: x = 2.3 + 1.3; Could be simplified to: x = 3.6; David Daney
