Difference between revisions of "Simulink/T-VEC Examples"

From T-VEC Wiki
Jump to: navigation, search
(Signal Ranges)
Line 9: Line 9:
  
 
===Signal Ranges===
 
===Signal Ranges===
 +
Signal ranges information defines the low bound and high bound values of input and output signals. This information is critical because test vector inputs are selected based on the range of the input signals. In addition, T-VEC analyzes the model using domain analysis starting from the signal range. The default input ranges are based on type definition. The min (low bound) and max (high bound) values often do not make sense. This especially true for floating-point numbers that have the following ranges:
 +
*32 bit float range: -3.4E+38 to 3.4E+38
 +
*64 bit float range: -1.7E+308 to 1.7E+308
 +
Selecting values in very large ranges degrades precision, for examples:
 +
:Y < X + 100  when X is 1.7E+308 is not meaningful
 +
:Value of 100 is lost in the noise of 1.7E+308
 +
 +
Default ranges for floating point types set to reduce problems
 +
*single (32 bit float) range: -1.0e4..1.0e4
 +
*double (64 bit float) range: -1.0e12..1.0e12
 +
Other types have default ranges
 +
*int8: -128..127 int16: -32768..32767
 +
*unint8: 0..255 uint16: 0..65535
 +
*int32: -2147483648.. 2147483647
 +
*uint32: 0..4294967295
 +
The default type ranges can be modified or specific signal ranges can be defined by the users. These operations can be performed using the Signal Range Editor GUI or they can be done to the underlying file (XML) using an editor.
 +
 +
To create a default, run the SL2TVEC translation once, then use the Signal Range Editor to appropriately set the signal ranges.
  
 
===Test Sequences===
 
===Test Sequences===

Revision as of 16:23, 14 February 2007