All compute nodes are shared among multiple users and jobs. Resources like GPUs and compute cores are never shared. In some cases, e.g. for benchmarking, exclusive access to the compute node can be desired. This can be achieved by using the Slurm parameter --exclusive.
Setting --exclusive only makes sure that there will be no other jobs running on your requested nodes. It will not automatically give you access to all resources of the node without explicitly requesting them. This means you still have to specify your desired number of GPUs via the gres parameter and you still have to request all cores of a node if you need them via --ntasks or --cpus-per-task.
Do not use --exclusive by default, as it denies other users to use the remaining ressources on a node you didn't request. If your account has accounting limits, then using --exclusive might eat up your accounting quota faster than expected, because all ressources of a node (CPU-hours, GPU-hours) will be charged.
