Implicit conversion from float -8.125 to int loses precision