Dose Profiler simulation @ GSI

Working @ GSI

• First of all one has to log onto the gateway (the kronos cluster is not directly accessible from outside)
  • ssh -X -Y -C username@lx-pool.gsi.de
• On the lx-pool cluster one has some space (~100 GB) in a given home under /u/username. Such space is limited and is not 'seen' by the kronos machines.
• To submit the jobs and use the computing resources, one has to log onto the kronos cluster:
  • ssh -X -Y -C username@kronos.hpc.gsi.de
• On kronos one has access to the nyx filesystem
  • /lustre/nyx/bio/first/
  • Such filesystem is accessible from all the kronos nodes, used to run the jobs using slurm
  • Eg.: vincenzo and alessio have their homes under /lustre/nyx/bio/first/vpatera and /lustre/nyx/bio/first/asarti
• To setup the software tools needed to run fluka or root on the kronos cluster use:

  • export FLUKA=/lustre/nyx/bio/first/vpatera/local/fluka_2017
    export FLUPRO=$FLUKAGSI
    export FLUTIL=$FLUKAGSI/flutil
    export FLUFOR=gfortran

    export ROOTSYS=/cvmfs/it.gsi.de/root/v6-06-06
    export LD_LIBRARY_PATH=$ROOTSYS/lib/root:$ROOTSYS/lib

  • To use root:
    • module use /cvmfs/it.gsi.de/modulefiles
    • module use /cvmfs/it.gsi.de/root/v6-06-06
• To submit jobs using the slurm queue one has to have the slurm account.
  • A slurm tutorial can be found here
  • To start with slurm first of all check if you can the commandsbatch /lustre/nyx/bio/first/asarti/test.sh
  • Once you know better what you are doing, check how to submit array of jobs havin a look at: /lustre/nyx/bio/first/asarti/DPSimulation/ScriptSubmission_v0.sh

-- Alessio Sarti - 2018-10-03

Comments

<--/commentPlugin-->