Last modified 4 years ago Last modified on 12/18/13 12:07:08

Cosmo cluster

Other HPC resources in CP3.


  • Nodes: 1 front-end + 12 nodes in the group "planet" + 4 nodes in "exoplanet/zodiacus" + 4 nodes in "exoplanet/hydra"
  • Processors/planet: 2 Quad-Cores AMD Opteron
  • Processors/exoplanet/zodiacus: 2 Eight-Cores AMD Opteron
  • Processors/exoplanet/hydra: 2 Eight-Cores-Hyper-threaded Intel Xeon E5
  • RAM/thread: 2GB (planet: 16GB, exoplanet/zodiacus: 32GB, exoplanet/hydra: 64GB)
  • Disk/planet: 1T
  • Disk/exoplanet: 2T
  • Connection: ethernet Gigabit stripped over 2
  • Filesystem: 25 TB parallel fhgfs + 800 GB nfs (home directories)
  • OS: Centos 6.5
  • Software: Open Source only
  • Queue System: Torque, support for OpenMPI and OpenMP


Request login from Christophe Ringeval, then from our network:

    ssh <login>@cosmo


    ssh <login>

Tips for SSH.

Queues and submitting jobs

Queues listing and specifications are retrieved with the command

  qstat -q

Running and queued jobs

  qstat -a

Submit jobs

  qsub <script_filename>

Example scripts are provided in your home directory.