As @Carles wrote, you can only use one computer, if you want, run both the controller ( slurmctld ) and the working ( slurmd ) daemon.
If you want to test some configurations and observe the behavior of Slurm, you can even run several working daemons on the same machine to simulate a larger cluster using the -N <hostname> parameter.
If you want to perform some calculations, you can run the controller and the working daemon on the same node. If you want the system to still be responsive, just configure Slurm to assume that the system has 1 core and 2 GB less RAM than it actually does, to leave room for the OS and the Slurm daemons.
As an additional note, the pages you specify in your question correspond to a very old version of Slurm. A newer version of the documentation is available on the Schedmd website .
source share