On supported architectures, change the rounding mode of the floating point arithmetics (to nearest, up, down, to zero) or change the precision of the arithmetical operations (single, double, double extended). Experimentally test the properties of the floating point arithmetics.
Check some properties of floating point arithmetics: the rounding mode and the machine epsilon, as seen by Octave's interpreted code (ROUND and MACHEPS, respectively) and, if requested, as seen by BLA