Skip to content

ParaView

ParaView is an open-source multiple-platform application for interactive, scientific visualization.

Homepage Version(s): 5.5.2

ParaView requires an environment module

In order to use ParaView, you must first load the appropriate environment module: 

module load gnu openmpi paraview

ParaView is a tool for developing scalable parallel processing tools with an emphasis on distributed memory implementations. The program is capable of rapidly building visualizations for data analysis using qualitative and quantitative approaches.

Data analysis and exploration in ParaView can be done either interactively in a 3D display or in non-interactive mode by using the built-in batch processing capabilities.

ParaView is best suited for exploring very large data files that are too large to fit in the memory of single node. It is also good for generating visualizations and simulations.

There are three basic components to ParaView:

  • Data server - data processing
  • Render server - data rendering
  • Client interface - for user interaction

Running via Open OnDemand#

The recommended way to interactively run ParaView is via our ParaView Open OnDemand app. Refer to our OOD documentation to get started. You can develop your ParaView scripts in the graphical IDE, and then take advantage of Slurm scheduler as described below.

Submitting ParaView jobs to the HPC#

To take full advantage of our HPC cluster with ParaView, you can schedule jobs to run non-interactively on the cluster. An example submit script is below:

1
2
3
4
5
6
7
8
9
#!/bin/bash
#SBATCH -A genacc_q
#SBATCH -n 8

# Load necessary modules
module load gnu openmpi paraview

# Run paraview 
pvbatch --mpi my_pv_script.py

Advanced usage#

In this configuration, the client GUI runs on the user's computer and uses SSH tunnelling as described below to connect to a server running on an HPC compute node. This configuration combines the data/render servers into one unit, which can then be run in parallel on multiple compute nodes via MPI.

Warning

The following technique requires connectivity to the FSU network, either on-campus or via the FSU VPN 

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
# From the login cluster, start an interactive job... 
# (if your research group has purchased dedicated resources, substitute your Slurm account for 'backfill')
[USERNAME@hpc-login.rcc.fsu.edu ~]$ srun -n 2 -A backfill -t 8:00:00 --pty /bin/bash

# When your job starts...
# (take note of the hostname before the prompt; in this example, it is 'hpc-i35-1-1')
[USERNAME@hpc-i35-1-1 ~]$ pvserver --mpi
Waiting for client...
Connection URL: cs://hpc-i35-1-1.local:11111
Accepting connection(s): hpc-i35-1-1.local:11111

On your personal computer, create an SSH tunnel:

1
$ ssh -Y -N -L localhost:11111:hpc-i35-1-1:11111 USERNAME@hpc-login.rcc.fsu.edu

You will need to leave this session running for the duration of your job.

Open the ParaView IDE on your personal computer, and then do the following:

  1. From the "File" drop-down menu, select "Connect", and click then "Add Server" button. Fill out the form as follows:
    1. Name: Anything you want (e.g., "FSU ParaView Job")
    2. Server "Clint/Server"
    3. Host: "localhost"
    4. Port: "11111"
  2. Click the "Configure" button.
  3. Set the Startup Type to "Manual".
  4. Click the "Save" button.

Once the configuration is saved, click the "Connect" button. You can verify you are connected by selecting "Memory Inspector" from the "View" dropdown menu (View -> Memory Inspector). It should show both your client GUI and the server process.