You were both right. The error lied in the softwares I was using: wolfram mathematica and libreoffice calc. It seems that for floating numbers greater than 1, the number of digits after decimal point is rounded to six, hence the difference. There is no problem for amounts lower than 1.

Thanks.