It is possible estimate the computational cost of Simulink blocks before embed on the hardware?

In the development by functional blocks in Simulink, is it possible to estimate the computational cost of hardware (ram, processing, etc) before embed the code on the hardware (for any hardware, arduino, raspberry pi, and others)? Estimate the cost that each block will use and estimate the cost of the entire system.

Thanks!

Answer 1

[Shaunak]

Hi Victor,

In Simulink, diagnostics can be performed on the full model to assess computational aspects before deploying to hardware like Arduino, Raspberry Pi, or other embedded platforms. While there is not a direct per-block computational cost estimate, several built-in tools are available for analyzing the overall model.

The ‘sldiagnostics’ command can be used to obtain computation statistics for the entire model. Additionally, the “Simulink Profiler” helps you identify bottlenecks and measure simulation performance, while the “Performance Advisor” provides recommendations for improving model efficiency.

Kindly refer to the following documentation for more information on profiling and performance analysis in Simulink:

https://www.mathworks.com/help/simulink/slref/introduction-to-profiling-models.html

Kindly refer to the following MATLAB Answers thread for further reference:

https://www.mathworks.com/matlabcentral/answers/378307-does-matlab-allow-us-to-analyse-the-computational-cost-over-a-cosimulation-simulink-gt

Hope this helps!