[geth03]

hi all,

i have a windows server with two additional virtual machines on it.
i am using the server with two other co-workers.

when i start a simulation with the virtual machines, the used cores are not fixed, they change the whole time, i attached two screenshots for two different times.

the problem is, when i use nearly all cores, because the simulations do not use the same cores, it slows down the computation time.
i think thats because when the cores are switched, the simulation waits until it can use the core, which will be free when calculation is done there.

i already used mpirun options like --bind-to core or --bind-to socket and -cpu-set but no success so far.
i hope anybody can help in this regard.