In 5G Toolbox the system-level simulation is very slow

I'm using the PHY abstraction in The System-Level Simulation, yet the code runs very slowly (several hours to process a few frames on an ordinary Dell laptop). I have multiple cells (3) and multiple users per cell (6) , so that has an obvious impact on the run-time, but these are small numbers. Are there any tricks to speed this up? Our code is similar to provided examples such as "NR Cell Performance with Downlink MU-MIMO" and "NR Intercell Interference Modeling." I tried setting enableTraces to false, but that didn't help. I see there is a variable 'numMetricsSteps' that the code states affects simulaiton time, but I'm not sure if lowering it it will reduce accuracy?.Any other suggestions for speed-up?

Answer 1

[Shubham]

Hey Laurence,

Refer to the following suggestions for improving simulation time:

https://www.mathworks.com/help/simulink/ug/speed-up-simulation.html

https://www.mathworks.com/company/technical-articles/improving-simulation-performance-in-simulink.html

You may also use profiling tools to your advantage:

https://www.mathworks.com/help/simulink/performance.html

https://www.mathworks.com/help/simulink/slref/simulinkprofiler.html

https://www.mathworks.com/help/slcoverage/analyze-coverage-and-view-results.html

Although these are general suggestions but I hope these would help you out.

[Laurence Mailaender]

Hi Shubham. Thanks, but your tips are about simulink. I'm asking about a particular system-level tool from the 5G Toolbox. See https://www.mathworks.com/help/5g/end-to-end-simulation.html. I'm running a modified version of this example: https://www.mathworks.com/help/5g/ug/nr-cell-performance-with-downlink-mu-mimo.html. Are you familiar with this kind of simulation?