How can I check the performance of a simulink model, which should be compiled for embedded hardware later?

Hi,

I have a question regarding the performance of a simulink model. I am developing an algorithm, which should be compiled into c-code for embedded hardware later. I developed two different models, both of them have the same task. Now I want to estimate, which one is more efficient regarding the cpu load. Is there any possibility to do this? I already tested the performance advisor and reading the metadata of a simulation ( i.e. the executionElapsedWallTime). But this does not really help, because the execution time can also rise when the computer does other things in the background.

Is there another option, which might help me?

Answer 1

[Mark McBroom]

You can measure execution time of the generated code on the host computer or on the targer processor using SIL or PIL mode and the execution profile feature: https://www.mathworks.com/help/ecoder/ug/execution-time-profiling-for-sil-and-pil.html