Simulink - Low Preformance after Simulation

1 vue (au cours des 30 derniers jours)
Maik Friese
Maik Friese le 1 Nov 2021
Modifié(e) : Simar le 16 Mai 2024
Hello Community,
I have a big Simulink - Model with estimated 1000 Inputs and 2000 Outputs. The work with Simulink is delayed but acceptable. After Simulation, the Preformance of Simulink is very delayed. Working is not acceptable anymore. The CPU and GPU capacity are far away from 100%.
Please, it's possible the gain the Simulink Preformance before Simulation?
Best Regards
Maik

Réponses (1)

Simar
Simar le 16 Mai 2024
Modifié(e) : Simar le 16 Mai 2024
Hi Maik,
I understand that you are facing issues with a large Simulink model, especially after simulations, despite low CPU and GPU usage and are seeking advice on improving Simulink's performance before simulation. To enhance performance for large models, consider these streamlined strategies:
  • Model Structure: Break down the large model into smaller subsystems using Model Referencing for more efficient processing.
  • Solver Settings: Opt for Fixed-Step Solvers as they offer better performance compared to variable-step solvers. Also try adjusting Step Size as a larger step size can speed up simulations.
  • Execution Efficiency: Employ Accelerator and Rapid Accelerator modes to compile parts of the model to machine code, speeding up execution.
  • Data Handling: Log only essential signals and use efficient formats like the dataset logging format. For frequently adjusted parameters, use real-time tuning methods to avoid recompiling the entire model.
  • Computational Resources: Can use "Parallel Computing Toolbox" for running simulations in parallel.
  • Optimization Tools: Use "Simulink Profiler" to Identify and address bottlenecks in model execution
Applying these focused strategies can significantly improve the handling and performance of large Simulink models, making simulation tasks more efficient.
Please refer to the following documentation links
Hope it helps!
Best Regards,
Simar

Catégories

En savoir plus sur Multicore Processor Targets dans Help Center et File Exchange

Produits


Version

R2020b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by