Floating Point Rounding Errors In Algebraic Processes