Re:GCC optimized code gives strange floating point results
by caf on Sunday February 26, 2017 @09:50PM
(#13741474)
The -fexcess-precision=standard option, which is enabled by -std=c99 (requesting ICO C99 compliance), forces excess precision to be rounded away on all assignments and casts to narrower precision types.
See https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html#ind...