MATLAB: Does the code generation of elementary math functions not respect single-precision floating point data type

abscodedoubleEmbedded Coderfunctionsgenerationmathmatlab codersimulink codersingle

I am using the ABS block inside my Simulink model with input datatype as 'single'. The outport of the model is set to inherit input data type. When I generate code from the model, I see the following:

out_sig_single = (F32)fabs(in_sig_single);

Even though the results would be computationally correct, the generated code uses double precision floating point numbers for operation and then uses cast operator to cast it to a single precision floating point data type. My hardware only supports single precision Math.

Best Answer

This feature of code generation for elementary math functions (such as ABS block) is not supported, as of MATLAB R2019a.

The workaround is to change the standard math library by going to 'Configuration Parameters > All Parameters' and searching for "Standard math library". Then please change this to "C99 (ISO)".

In older releases, you may find this option in 'Configuration Parameters > Code Generation (or Real-Time Workshop) > Interface' and possibly with a different name: "Target function library".

Before changing the library, please make sure to delete the 'slprj' folder present in the current directory.

Related Solutions

MATLAB: Do I get a ‘floating-point support’ error when Generating code for a S-function in Real-Time Workshop Embedded Coder 5.4 (R2009b)

This error is thrown when generating code for an integer model that contains a S-function with a double precision parameter that is defined as tunable and when the 'support floating point numbers' option on the Configuration Parameters > Real-Time Workshop > Interface page is not enabled.

The parameter 'P1' of block 'ExampleModel1/ExampleSFunction' is a floating point type. Consider changing the data type or selecting the 'support floating point numbers' option on the Configuration Parameters > Real-Time Workshop > Interface page.

By default, the s-function parameters are set to be tunable unless you specify that you don’t want them to be tunable using the following macro:

 ssSetSFcnParamTunable(S, 0, SS_PRM_NOT_TUNABLE);

You can find relevant information in the documentation page titled "Make a block parameter tunable"

http://www.mathworks.com/access/helpdesk/help/toolbox/simulink/sfg/sssetsfcnparamtunable.html

If you define a parameter as tunable, you will be able to change the value during simulation and run-time of the generated code and therefore the parameter will be included in the generated code. If the value is double and the rest of the model is integer, the parameter will need to be cast to integer in the generated code. The ‘floating-point support error’ is indicating that this parameter is double and that you should either turn on floating-point support or convert the parameter to integer. In this case, integer-only code and a non-tunable parameter are preferred, so changing the s-function parameter to not tunable resolves this issue.

Please see the link below for more information on support of floating-point numbers.

http://www.mathworks.com/help/releases/R2009a/toolbox/rtw/index.html?/help/releases/R2009a/toolbox/rtw/ref/bq259j_-1.html

MATLAB: Is code generation behavior different in 2018b for the “max” function

In 2018b, the default standard math library changed to "C99 (ISO)", hence the code generator makes the call to the "fmaxf" function which resides in the same library.

To return to the earlier code generation behavior, please change the standard math library to "C89/90 (ANSI)" in the following manner:

>> cfg = coder.config('dll');
>> cfg.TargetLangStandard = 'C89/C90 (ANSI)';
>> codegen myScript -config cfg -args {1 2 3} % replace the script name and args list accordingly

If you are using the MATLAB Coder App, this option will be found in the Generate Code step under More Settings > Custom Code > Code Replacement Libraries > Standard math library.

Best Answer

Related Solutions

MATLAB: Do I get a ‘floating-point support’ error when Generating code for a S-function in Real-Time Workshop Embedded Coder 5.4 (R2009b)

MATLAB: Is code generation behavior different in 2018b for the “max” function

Related Question