Our code gets little benefit from using virtual cores (hyperthreading), so when we run with mpiexec on an 8 real plus 8 virtual machine, we would like to be certain that it uses only the 8 real cores.
Is there a way to do this with openmpi? Thx....John
