Dear NSEM community,
I would like to ask for your advice regarding some irregularities I encountered in my simulations.
The setup: I am running NSEM simulations on a shared Windows machine (one PC, multiple user accounts). At the same time, colleagues often run CST simulations on the same machine. Unfortunately, I was not able to run my simulations in a situation where I was the only user.
The issue:
-
Normally, filling the impedance matrix takes about 290s.
-
In my case, the filling process stopped at some point until I pressed Return in the console. Afterwards, the simulation continued with the next frequency point.
-
A second, longer interruption happened overnight, visible only the next morning, so the amount of time is higher.
-
These interruptions occurred across different simulations with the same data/framework.
My questions:
-
Could running CST in parallel interfere with NSEM (e.g., by setting registers also used by NSEM)?
-
Could the issue be related to the machine itself, or still a user-side problem (CAD, Python code, etc.)?
-
What about the results: are the data from the interrupted runs (e.g., 9th and 13th iteration) defective, or are they still valid except for the interrupted points? Should I always restart the full run in such cases?
Screenshots of the interruptions are following for reference.
Thank you very much for your help and insights.
Best regards
Michael
marginal information
Simulation:
1 to 2 GHz
28 Frequency steps
dence solver
quad9
7102 number of unknown
Both objects in one block
goal: simulate the scattered reflection of the second object as S11 (over distance after post processing)
python code
If heare are syntaxerrors than only becouse of deleting not necessary parts to make it more readable. In general it works. For hints, I’m still open.
pythoncode.txt (913 Bytes)
Sim Log file .OUT
SIM_LOG.txt (48.9 KB)

