After compiling with CMake, setting the same random generator seed across simulation runs produces different precise spike times.
It has been confirmed that model behavior is still statistically valid using the methods of Tiddia et al. 2022.
This leads us to believe the issue is related to the random numbers generators, however this issue is NOT present when compiling with Autotools so it is certainly related to CMake itself.
For now compiling with Autotools will ensure having reproducible, precise spike timings.
After compiling with CMake, setting the same random generator seed across simulation runs produces different precise spike times. It has been confirmed that model behavior is still statistically valid using the methods of Tiddia et al. 2022.
This leads us to believe the issue is related to the random numbers generators, however this issue is NOT present when compiling with Autotools so it is certainly related to CMake itself.
For now compiling with Autotools will ensure having reproducible, precise spike timings.