Hi,
We are currently working with EP FMU and we encountered two issues:
The first one is about simulation time, the simulation was in quadratic time complexity O(n) with n the number of variable exchanged by the FMU. We found that SetReal and GetReal functions in the FMU are in linear time complexity O(n). We changed them by removing a verification on variables to get a constant time complexity. We are not sure what the purpose of the verification we removed is and how it could affect the simulation.
this is "SetReal" before the change :

this is "SetReal" after the change :

The second issue is the number of variable the FMU can handle. When we tried to use the FMU with more than 1100 inputs or 1100 outputs, EP seems to crash at the first step of the simulation. The crash doesn’t seem to be related to the total number of variable exchanged with the FMU, but only with the number of inputs or outputs separately. This should means that the issue for output and input is the same but they still are independent. For example, an FMU with few inputs and 1200+ outputs will crash, same for few outputs and 1200+ inputs, but an FMU with about 700 inputs and 700 outputs won’t crash. When we launch the simulation, all the warming up works but at the start of computation the EP windows close. What is strange is that Matlab or Python keeps running using the FMU but nothing is written neither in output files nor in log files. We suspect the issue to be a memory error related to sockets but we can’t say for sure, and we didn’t find any fix for this. Do you know where the issue could come from and how to fix it? We gave you some of the IDF we used for our tests as attached files. To create our FMU we took a small IDF then we added useless variables as inputs and outputs in order to test the communication process without changing the simulation of EP.
Thank you,
Quentin DE NAPOLI COCCI