This cluster consists of 34 nodes connected via 56Gb FDR Infiniband. Each node has two four-core 3.3 GHz Intel Xeon E5-2643 processors and 256GB of system memory.
Accessing the System
Note: Not all projects have access to the Intel cluster. Job submissions to Slurm may be rejected even if access to the front-end node is authorized.
Running on the cluster first requires connecting to its front end node
rsa35. This machine is accessible from the landing pads.
By default Slurm will assign 16 processes to each node. The 2x factor is the result of hyperthreading being enabled. Some applications may benefit from hyperthreading, others will not. Initial testing with Fluent indicates that running one process per physical core yields the best performance.
Passing the '--bind-to-core' option to OpenMPI will specify process affinity to cores, and along with Slurm options '-N', to specify the number of nodes, and '-n', to specify the number of processes, the physical cores will each run a single process. For example passing '-N 2 -n 16' to Slurm and '--bind-to-core' to mpirun will result in 16 processes running on 16 cores on two nodes.
Alternatively, passing '-c 2' to srun will assign two cores per process and will prevent execution of more than 8 processes per node.
MVAPICH2 and OpenMPI compiler wrappers are available via the 'mpi' modules. Please refer to Modules for use of modules and their interactions with Slurm.
Submitting and Managing Jobs
Please see Slurm.
|Name||Time Limit (hr)||Max Nodes|