Floating-point math is hard, and compilers will exploit every language loophole to make our FP calculations go faster, sometimes with surprising results.
Floating-point math is hard, and compilers will exploit every language loophole to make our FP calculations go faster, sometimes with surprising results.