Hi, All:

Does anyone know how to control node selection when jobs were submitted from Windows GUI to Linux cluster via MS6.1 Gateway?

Note: our linux cluster does not have any queuing system.

I knew file \\\$MSHOME/share/data/machines.LINUX used to control node selection for MS prior to 6.0 version when HPMPI was used to start the MPI job.

After Intel MPI was used to start MPI job since MS6.0, I am not sure how to allocate the job to any internal compute node instead of the default head node of a Linux cluster

when jobs were submitted from Windows GUI to Linux cluster via MS6.1 Gateway.

Any help will be appreciated!

Jun Wang