Floating point math is not associative, (x+y)+z != x+(y+z). Most of the time the difference is small, but in some cases not (I.E. x = 1e30, y = -1e30 and z = 1,

Floating point math is not associative, (x+y)+z != x+(y+z). Most of the time the difference is small, but in some cases not (I.E. x = 1e30, y = -1e30 and z = 1, it is 1 in the former case, 0 in the latter). Modern compilers will reorder operations depending on a whole host of factors, and this means in general unless you really go out of your way, floating point math may not even be deterministic! I've seen this happen during certain optimization algorithms, completely different results on different machines even with the same random seed and confirmed same random sequence of numbers, it was because of the floating point reordering! https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.h...