Hi,
I know something defferent!
You could run succesfully on a single-CPU 8 cores, but you may not be able to run on 2 CPUs x 4 cores each) because of MPI installation.It depends your network and your HPC system as well.
If you have to carry out your simulation by submitting a job, I would prefer using 2CPUs x 16, rather than 4 CPUS x 8cores. One of reasons for this: 4x8cores usually takes longer time to be accepted.
Therefore, please try first. After that you can find the answer!
Hope this helps
Regards
Huy