C Compiler fast math

Back

-ffast-math enables certain unsafe floating point optimizations, it is a collection of different flags. These optimization turn off IEEE 754 cases like NaN and infinity. It is usually better to enable specific flags based on the work load like -funsafe-math-optimizations -fcx-limited-range

Further Reading

https://gcc.gnu.org/wiki/FloatingPointMath