MATLAB: RAM/ROM memory metrics for Embedded Coder

Embedded Coder

I am generating code from a Simulink model using Embedded Coder. Is there any way that Simulink can measure/predict the amount of ROM/RAM that the application will need?

Best Answer

Listed below are several documentation links pertaining to retrieving code metrics from Embedded Coder:

Generating Static Code Metrics

https://www.mathworks.com/help/rtw/ref/static-code-metrics.html

https://www.mathworks.com/help/ecoder/ug/generate-a-static-code-metrics-report.html

Memory Usage Optimization

https://www.mathworks.com/help/ecoder/memory-usage.html

Performance Evaluation of Generated Code

https://www.mathworks.com/help/ecoder/examples/evaluating-the-generated-code.html

Additionally, under the "Summary" section of the "Code Generation Report", you can find an entry titled "Memory Information" under the "Code Information" table. This table entry would be present if the "Static Code Metrics" Configuration Parameter is enabled.

The "Memory Information" entry would report the "Global Memory" as well as the "Maximum Stack" size of the generated code. Since true memory usage is platform and compiler dependent, an exact memory usage metric would not be available solely from code generation. The exact number would have to be retrieved from the execution of compiled code through third-party or compiler packaged utilities.

The last documentation link highlights an example of using a utility from NXP to determine certain runtime metrics such as RAM and ROM usage.

Related Solutions

MATLAB: What embedded compilers and Integrated Development Environments (IDE) does Embedded Coder support

Embedded Coder generates ANSI/ISO C/C++ code and thus supports all embedded compilers and IDEs that support C/C++ code. If portable ANSI/ISO C/C++ that can be called by any scheduler is not sufficient, Embedded Coder has a variety of options and features that end users can leverage to generate target specific code for their particular compiler environment. A sample of these options and features are listed below.

Optimize Generated Code by Developing and Using Code Replacement Libraries - MATLAB and Simulink:

https://www.mathworks.com/help/ecoder/examples/optimize-generated-code-by-developing-and-using-code-replacement-libraries-mc.html

https://www.mathworks.com/help/ecoder/examples/optimize-generated-code-by-developing-and-using-code-replacement-libraries-sc.html

Custom Simulink Real-Time drivers (R2017b and earlier releases):

https://www.mathworks.com/help/releases/R2017b/xpc/driver/introduction.html

Memory/Pragma Sections:

https://www.mathworks.com/help/ecoder/ug/control-data-and-function-placement-in-memory-by-inserting-pragmas.html

Legacy Code Integration using Legacy Code Tool:

https://www.mathworks.com/help/simulink/legacy-code-integration.html

Control generation of Function Prototypes:

https://www.mathworks.com/help/ecoder/ug/support-c-function-prototype-control.html

Code Generation from MATLAB and Simulink Code:

https://www.mathworks.com/help/ecoder/c-c-code-generation-from-matlab-code.html

https://www.mathworks.com/help/ecoder/c-c-code-generation-from-simulink-models.html

MATLAB Central contains several examples that may also be relevant to your needs:

<https://www.mathworks.com/matlabcentral/fileexchange/?term=Embedded+coder+type%3A%22Example%22&utf8=%E2%9C%93>

Note that MathWorks does not guarantee or warrant the use or content of MATLAB Central submissions. Any questions, issues, or complaints should be directed to the contributing author.

In addition to ANSI C/C++ code generation and target customization options mentioned above, MathWorks and third party link and target products offer built-in integration support for several target, compiler, and RTOS combinations. These products provide device driver blocks, target optimized code, build environment automation, Processor-in-Loop testing, and other features.

You can find all the relevant details related to Embedded Coder in the below link:

https://www.mathworks.com/products/embedded-coder.html

Finally, MathWorks consulting and a number of third parties service providers are able to create custom solutions for your target environment needs.

MATLAB: How to insert macro before and after the function corresponding to atomic subsystem in the generated code

This can be done using memory section. 
1. Create a class data package 'mypkg' using the following steps:

View the +SimulinkDemos data class package in the folder _matlabroot_/toolbox/simulink/simdemos/dataclasses. This package contains predefined data classes.

Copy the folder to the location where you want to define your data classes.

Rename the folder +mypkg and add its parent folder to the MATLAB path.
2. Run the following command to open the custom storage class designer:
>> cscdesigner('mypkg')
3. Under "Memory Section", create a new Definition, 'myMS'. 
4. Selecting this, in the "Pre-memory-section pragma" field, enter "#define a 1" without the quotes. And in the "Post-memory-section pragma" field, enter "#define end 1" without the quotes.
5. Click OK and restart MATLAB to reflect the change.
6. In your model, go to Configuration Parameters -> Code Generation -> Memory Sections and then select the 'mypkg' package from "Package" drop-down list.
7. Right click the atomic subsystem and select "Block Parameters (Subsystem)", under "Code Generation" select 'myMS' from "Memory section for execution functions" drop-down list
8. Click OK and generate code for the model.
More information can be found in the following documentation page:
<https://www.mathworks.com/help/releases/R2017b/ecoder/ug/control-data-and-function-placement-in-memory-by-inserting-pragmas.html https://www.mathworks.com/help/ecoder/ug/control-data-and-function-placement-in-memory-by-inserting-pragmas.html>

Best Answer

Related Solutions

MATLAB: What embedded compilers and Integrated Development Environments (IDE) does Embedded Coder support

MATLAB: How to insert macro before and after the function corresponding to atomic subsystem in the generated code

Related Question