There are two methods to execute serial jobs in pratyush

1 . Using Python code for submission of serial job in compute node.
     The serial codes are submitted to PBS with aprun command and for single      compute node.The jobs are forwarded to the compute nodes instead of      running on MOM nodes.


Python script for submission serial job in compute node .
Please Refer Following files (Use wordpad for proper indentation) for serial job submission on PRATYUSH using python script.
sub_python.pbs
extract.py
sub.pbs

2. MAMU Nodes

MAMU(Multiple Application Multiple Users) nodes are used for serial workload and intra-node message passing interface (MPI).
Jobs are directly placed on the node by submitting jobs into PBS(Scheduler).
These nodes cannot be used to place jobs on compute nodes as they are not managed by ALPS.

On MAMU nodes for MPI jobs ,please ensure you have loaded snplauncher(/opt/pbs/default/bin/mpiexec).

MAMU Node details -:
        Number of nodes = 4
        Number of core/node = 72
        Total Number of cores = 288

Please find the steps to submit jobs into MAMU nodes -:
       Maximum ncpus = 288

Sample Job Script -:
#!/bin/bash
#PBS -l select=1:ncpus=1:vntype=cray_mamu
#PBS -q serial
module load cray-snplauncher/7.6.3
mpiexec -n[-np] 6 ./application

Note -: The jobs submitted to "serial" queue won't be forwarded to workq.