Specifying Data Type for Code Generation with IQmath Library

I am using the Embedded Coder Support Package for TI C2000 Processors. How do I specify the output data type for a Simulink block to be in the 'Q' format, as used in the TI IQmath library?

Answer 1

[MathWorks Support Team]

The correct way to configure a block to use the Q format is to specify a fixed-point output data type that matches the particular Q format in use. For example, to configure a block for a Q23 representation, specify the data type as "fixdt(1, 32, 23)" under Block Parameters > Signal Attributes > Output data type. This indicates a signed 32-bit fixed-point number with 23 fraction bits, which matches the Q23 representation. This is described on the following documentation page, under the heading "Q Format Notation":

https://www.mathworks.com/help/releases/R2023a/ti-c2000/ug/using-the-iqmath-library.html#mw_ee4b5104-d15d-4973-ba63-0789ce3c57a0

Please see the documentation page of "fixdt" for more information on creating Fixed-Point Numbers in Simulink:

https://www.mathworks.com/help/releases/R2023a/simulink/slref/simulink.numerictype.fixdt.html

Additionally, please refer to the user's guide for the 

C28x IQmath Library - A Virtual Floating Point Engine

, Literature Number SPRC087, available at the Texas Instruments Web site for more information on the IQmath library.