MicroElec Results Linux vs Windows

Hello,

I have a project I created on windows concerning the electron emission from a tungsten dust particle with energies of the beta decay of tritium. I am using MicroElec and Opt4 to simulate this and produce good results on windows. Now i need to transfer my project to a linux cluster to simulate some heavy runs. I copy the project over and build it in linux with the same version of geant4 (11.3.0), i try to use settings to reflect the compiler of windows (i use Visual Studio Code to build my project), however i just get different results when i run on linux (more electrons detected leaving the dust particle, around 20% more so its a large error). i even set the random seed the same and make sure to use 1 thread on both to see what the issue is. With this, the starting energy is the same of course but on my output i see there are more secondaries being created on the linux run. Why could this be? I need to perfectly replicate the output of my windows