Hi NVIDIA Aerial Team,
We are trying to modify testCompGenTV_pusch.m configurations and generate as testVectors.
In our understandings, it seems that createPythonPackage.m has something to do with generating aerial mcore package.
However, we do not have the full instructions for generating and installing aerial mcore package.
Also, does the matlab version has to be 2023a for executing createPythonPackage.m?
What if we run createPythonPackage.m on different matlab ver., i.e. 2025a, and install dependent python environment on the same target machine, is the testVector generated able to run properly with pusch_rx_multi_pipe/cuphy_ex_pusch_rx_multi_pipe tests?
Can you help with these problems?
Thanks.
Hi @Andrew.Kuo,
aerial_mcore is intended for developers who do not have licenses for the required Matlab toolbox licences. I assume that you have these licenses since you intend to re-generate the aerial_mcore.
If you are trying to generate new test vectors by modifying the testCompGenTV_pusch.m, you do not need to re-generate the aerial_mcore.
Also, does the matlab version has to be 2023a for executing createPythonPackage.m?
This should not be required but we have not tried it.
What if we run createPythonPackage.m on different matlab ver., i.e. 2025a, and install dependent python environment on the same target machine, is the testVector generated able to run properly with pusch_rx_multi_pipe/cuphy_ex_pusch_rx_multi_pipe tests?
As long as you match the aerial_mcore version with the Matlab runtime on your machine, it should work. If the testvectors are successfully generated cuphy pipelines will work. cuphy pipelines do not care about the Matlab version.
If you do not use aerial_mcore, you can directly generate the test vectors from Matlab as explained here .
Please also see here for some sample commands:
From within the 5GModel/nr_matlab/ directory in MATLAB:
startup
% Generate specific PUSCH test cases
testCompGenTV_pusch([7201, 7202], 'genTV')
% Or generate a full set
testCompGenTV_pusch('full', 'genTV')
% Or use the regression harness for PUSCH only
runRegression({'TestVector'}, {'pusch'}, 'full')
% Or for all channels
runRegression({'TestVector'}, {'allChannels'}, 'full')
Thank you.