mycluster 0.3.16

Utilities to support interacting with multiple HPC clusters

MyCluster

Master

Develop

Utilities to support interacting with multiple HPC clusters

Provides the ability to interact with the most popular HPC job scheduling systems using a single interface and enables the creation of job submission scripts. The system also tracks statistics of each job and records the hardware details of the compute nodes used.

Data is stored in a local database in ~/.mycluster

Tested with SGE and slurm (LSF and PBS/TORQUE support under development)

Installation - pip install MyCluster or git clone https://github.com/zenotech/MyCluster.git

Dependencies: ZODB, Fabric

Example usage

Register details

mycluster --firstname Fred --lastname Bloggs --email [e-mail]

List all queues

mycluster -q

Create job script

mycluster --create JOB_SCRIPT --jobqueue QUEUE --script SCRIPT --ntasks=TASKS --jobname=JOB_NAME 
          --project ACCOUNT_NAME --maxtime 12:00:00

Submit job

mycluster --submit JOB_SCRIPT

Delete job

mycluster --delete JOB_ID

Print job table

mycluster -p

Print help

mycluster --help

The SCRIPT to be executed by the JOB_SCRIPT can make use of the following predefined environment variables

export NUM_TASKS=
export TASKS_PER_NODE=
export THREADS_PER_TASK=
export NUM_NODES=

# OpenMP configuration
export OMP_NUM_THREADS=$THREADS_PER_TASK

# Default mpiexec commnads for each flavour of mpi
export OMPI_CMD="mpiexec -n $NUM_TASKS -npernode $TASKS_PER_NODE -bysocket -bind-to-socket"
export MVAPICH_CMD="mpiexec -n $NUM_TASKS -ppn $TASKS_PER_NODE -bind-to-socket"
export IMPI_CMD="mpiexec -n $NUM_TASKS -ppn $TASKS_PER_NODE"

In order to capture the relevant information it is recommended that the SCRIPT also exports the following environment variables

# Application name
export MYCLUSTER_APP_NAME=
# Data size that typifies application performance for this job (e.g number of points or number of cells)
export MYCLUSTER_APP_DATA=