Submission of job on HPC

Author:

Vinay K L

Copyright:

None

Date:

07/10/2022

Purpose

To set up a job on the HPC cluster.

1. To connect, use IP 172.27.1.152

2. Our institute has limited resource cluster with 14 nodes (1 login node) with 40 cores (Intel Xeon W series) on each nodes. Each node hosts a 196GB RAM and 2 nodes with Nvidia GPU.

3. By default each user can request maximum 80 cores at a given time. But on special request to IT department this can be extended.

4. There are 5 different time Q based on walltime. 1) ShortQ (24:00:00) 2) MidQ (48:00:00) 3) LongQ(96:00:00) 4) ELongQ(128:00:00) 5)iiserq (Unlimited, but without prior IT approval, job can be killed without notice)

5. PBS is used as job scheduler and handler and below is a PBS template for reference.

    #PBS -N Name_reflecting_in_queue
    #PBS -q iiser
    #PBS -l select=1:ncpus=40
    #PBS -l walltime=128:00:00
    #PBS -M vinaykl@students.iisertirupati.ac.in
    #PBS -V
    #PBS -o name_of_log.o
    #PBS -e name_of_log.err
   
    ncores=`cat $PBS_NODEFILE|wc -l`
   
   
    #source /home/app/intel/compilers_and_libraries/linux/bin/compilervars.sh intel64
    #source /home/app/intel/compilers_and_libraries/linux/mpi/bin64/mpivars.sh intel64
    #source /home/app/intel/compilers_and_libraries/linux/mkl/bin/mklvars.sh intel64
   
   cd $PBS_O_WORKDIR
   cd /path/to/your/current/working/directory
   
   command