3.3.4.3.1 Floating Point Numbers.

Floating point numbers are used wherever a single number has to be passed along. Usually models that perform characterization tasks use the float type. For instance, the arithmetic-model (see Section 3.6.1 on page ) allows for arbitrary calculus on floating point data.

The bound-float type is a sub-variant of the float type. The value at a port of bound-float must not be outside of its bounds which are given by the definition of the port. Bound floats are especially useful during TCAD experiments such as design of experiments, optimization, or statistical analysis, since these experiments make use of bounds and no further additional user interaction is necessary in this respect. It is to note that ports of type bound-float and float are compatible to each other as long as the value of a float ports is not outside of the constraints of a bound-float port.