[gmx-users] GROMACS with different gcc and FFT versions but one unique *tpr file

[Thomas Schlesier]

Dear all,
i have access to a cluster on which GROMACS is compiled with a different 
version of GCC and a different FFT libary (compared to the local machine).
Will this affect simulationns if i prepare the *.tpr on the local 
machine and run the simulation on the cluster and the local machine?

Sorry if this is a dumb question. I could imagine that the two 
simulations will be not numerical identical due to the different FFT 
libaries, but how strong this effect is and what else could happen i 
have no idea...

Greetings
Thomas