Check altmax behaviour for floats

[dlinssen]

The altmax parameter sets the upper boundary of the computational domain in units of the planet radius. It was originally implemented as an integer. We should verify whether it works correctly for float values too - specifically the potential round-off errors introduced.