MARVIN Cluster

From CAC Documentation wiki
Jump to navigation Jump to search

This is a private cluster.


  • Head node:
  • access modes: ssh
  • OpenHPC v1.3.8 with CentOS 7.8
  • 86 compute nodes with Dual 6-core X5670 CPUs @ 3 GHz, Hyperthreaded, 24 GB of RAM; 3 high memory nodes with 96 GB of RAM
  • Cluster Status: Ganglia.
  • "Why use a temporary directory?"
  • Submit HELP requests: Help OR by sending email to:

File Systems

Home Directories

  • Path: ~

User home directories are hosted on the head node and exported to the compute nodes via NFS. Unless special arrangements are made, data in user home directories are NOT backed up.

Globus Access

User home directories can be accessed on Globus. Under "File Manager" tab in Globus web GUI:

  1. Access "cac#marvin" endpoint.
  2. Authenticate using your CAC user name and password if prompted.


  • The cluster scheduler is Slurm. See Slurm documentation page for details.
  • Note: hyperthreading is enabled on the cluster, so Slurm considers each physical core to consist of two logical CPUs. See the slurm options section for using the correct options for your job.
  • Partitions:
Name Description Time Limit
viz 3 visualization Ensight Servers, each has 96GB RAM none
normal (default) all nodes except for those in viz queue none
all all cluster nodes none


Software Path Notes
*GNU Compilers 8.3.0 /opt/ohpc/pub/compiler/gcc/8.3.0 module load gnu8/8.3.0
*openmpi 3.1.4 /opt/ohpc/pub/mpi/openmpi3-gnu8/3.1.4 or /opt/ohpc/pub/mpi/openmpi3-intel/3.1.4 module load openmpi3/3.1.4
Intel Parallel Studio XE 2020.1.217 /opt/ohpc/pub/compiler/intel/2020/ module swap gnu8 intel/20.1.2017
Intel MPI 2020.1.217 /opt/ohpc/pub/compiler/intel/2020/compilers_and_libraries_2020.1.217/linux/mpi module load impi/2020.1.217
mvapich2 2.3.2 /opt/ohpc/pub/mpi/mvapich2-gnu/2.3.2 or /opt/ohpc/pub/mpi/mvapich2-intel/2.3.2 module load mvapich2/2.3.2
fftw 3.3.8 /opt/ohpc/pub/libs/gnu8/openmpi3/fftw/3.3.8 or /opt/ohpc/pub/libs/gnu8/mvapich2/fftw/3.3.8 module load fftw/3.3.8
hypre 2.18.1 (gnu) /opt/ohpc/pub/libs/gnu8/openmpi3/hypre/2.18.1, /opt/ohpc/pub/libs/gnu8/impi/hypre/2.18.1, /opt/ohpc/pub/libs/intel/openmpi3/hypre/2.18.1, or /opt/ohpc/pub/libs/intel/impi/hypre/2.18.1 module load hypre/2.18.1
  • Installed only on the head node and "viz" queue (compute-3-13 to compute-3-16)
ensight 10.0 /usr/local/CEI
  • Installed only on the head node and "viz" queue (compute-3-13 to compute-3-16)
VisIt 2.9.2 /opt/visit
  • Must default to use the following OpenMPI version for parallel visualization like this:
-bash-3.2$ mpi-selector --set openmpi-1.6.3-gcc-4.6.2
Defaults already exist; overwrite them? (y/N) y
-bash-3.2$ mpi-selector --query
  • Log out and log back in for the change to take effect.
Anaconda Python /opt/anaconda-python
  • Add the following line to your ~/.bashrc to use anaconda python:
export PATH="/opt/anaconda-python/bin:$PATH"

Quick Tutorial

The batch system treats each core of a node as a "virtual processor." That means the nodes keyword in batch scripts refers to the number of cores that are scheduled.

Select your default MPI

There are several versions of MPI on the Marvin cluster. Use the following commands to modify your default mpi.

  • mpi-selector --query -> shows your default mpi
  • mpi-selector --list -> shows all available mpi installations
  • mpi-selector --set <mpi installation> -> sets your default mpi, note, you will have to exit and log back in for this to take effect.

Running an MPI Job on the Whole Cluster

  • assuming /opt/openmpi/ is the default, the mpiexec options may change depending on your selected MPI.
  • First use showq to see how many cores are available. It may be less than 1152 if a node is down.
#PBS -l nodes=96:ppn=12    (note, this is PBS -l (small case L))
#PBS -N test
#PBS -j oe
#PBS -S /bin/bash

set -x

mpiexec --hostfile $PBS_NODEFILE <executable>     (where executable is the program you wish to run)

Running an MPI Job using 12 Tasks Per Node

Because the nodes have 12 physical cores, you may want to limit jobs to 12 tasks per node. The node file lists each node 1 time, so make a copy with each node listed 12 times, and hand that version to MPI.

#PBS -l nodes=48
#PBS -N test
#PBS -j oe
#PBS -S /bin/bash

set -x

# Construct a copy of the hostfile with only 12 entries per node.
# MPI can use this to run 12 tasks on each node.
uniq "$PBS_NODEFILE"|awk '{for(i=0;i<12;i+=1) print}'>nodefile.12way

# to Run 12-way on 4 nodes, we request 48 core to obtain 4 nodes
mpiexec --hostfile nodefile.12way ring -v

Running Many Copies of a Serial Job

In order to run 30 separate instances of the same program, use the scheduler's task array feature, through the "-t" option. The "nodes" parameter here refers to a core.

#PBS -l nodes=1  (note, this is PBS -l (small case L))
#PBS -t 30
#PBS -N test
#PBS -j oe
#PBS -S /bin/bash

set -x
echo Run my job.

When you start jobs this way, separate jobs will pile one-per-core onto nodes like a box of hamsters.

Running on a specific node

To run on a specific node use the host= option

#PBS -l host=compute-3-16       (note, this is PBS -l (small case L))
#PBS -N test
#PBS -j oe
#PBS -S /bin/bash

set -x
echo Run my job.

Running in the viz queue

To run in the viz queue use the -q option

#PBS -l nodes=1       (note, this is PBS -l (small case L))
#PBS -N test
#PBS -j oe
#PBS -S /bin/bash
#PBS -q viz

set -x
echo Run my job.

Running an interactive job

from the command line: qsub -l nodes=1 -I