Zeus HPC consist of 4 login nodes (zeus[1-4]), 2 file servers (zeus[5-6], 15 Tb each), 2 backup machines and following compute nodes separated in different categories:
- 18 nodes (zeus200-217) of Intel(R) Xeon(R) CPU E5-2440 (Sandybridge) — 12 cores per node & 48 Gb of RAM, & 2x Nvidia Kepler K20 GPUs
- 144 nodes of Intel(R) Xeon(R) CPU L5530 (Nehalem) — 8 cores per node & 48 Gb of RAM
- 1 node of Intel(R) Xeon(R) CPU E5-4620 (Sandybridge)– 32 cores per node & 512 Gb RAM
New (2016) Broadwell CPU based nodes:
- 56 nodes with 2 x Intel(R) Xeon(R) Broadwell CPUs E5-2683 v4 @ 2.10GHz (32 CPU-cores/node) 128Gb RAM/node: SLURM “–constraint=broadwell”
Amongst these 56 Broadwell nodes, there are
- 10 nodes with 2 x Intel(R) Xeon(R) Broadwell CPUs E5-2683 v4 @ 2.10GHz (32 CPU-cores/node) 128Gb RAM/node + 2x NVidia Kepler K80 GPUs/node: SLURM “–constraint=k80″
- 2 nodes with 2xIntel(R) Xeon(R) CPU E5-2683 v4 @ 2.10GHz (32 CPU-cores/node) 128Gb RAM + Xeon Phi coprocessor SE10/7120: SLURM “–constraint=phi”
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
In order to use Zeus HPC you need:
- have a user account (apply my e-mailing Alex Pedcenko — aa3025(at)coventry.ac.uk).
- connect to Zeus with ssh-session (see below the section “Connecting to HPC“)>
- write or upload your code/executable/files/model etc via ssh/sftp (i.e. WinSCP under Windows, “scp” under Linux/Mac)
- submit your job to one of the queues of Zeus (see below section SLURM)
In order to use compute nodes of Zeus to run your executable, code, or whatever software, you have to submit a job to the queue. This is done via SLURM queue manager, namely:
- writing slurm job submission script
- submitting the script to the queue (i.e. with “sbatch” command)
0 Comments.