Some modern processors have a higher floating-point than older models. To take advantage of the higher precision companies like Intel provide a C++ compiler. The C++ compiler includes a higher precision datatype so that the programmers can take advantage of this.
Floating points typically follow an IEEE standard that includes a sign bit for positive/negative; an exponent and a fraction. There is a degree of error in the values the longer the fraction is.
Decimals differ from this as they consist of a fixed range of values before and after the decimal point. Hence these tend to be used for financial were as in science a degree of error is normally anticipated. The error is decreased with higher precision.
When doing calculations different rounding methods is used in different countries and also in different sectors. So tax is rounded down in the UK and up in the USA for the 2 decimal places. However, in maths, I was taught to round up from 5 or down for less than 5 (some time ago), ‘midpoint rounding, away from zero’. I.e. 4.65 would be 4.7 to 1 dp. In other places, other methods are used.
The numbering system is slightly different in other countries as well. In Germany, the decimal place is a comma ‘,’ in the UK and USA the decimal place/point. Which is the same as the full-stop character or period if you’re in the USA.