Visualisation Using ParaView

ParaView

• Open-source, cross-platform data analysis and visualisation application
• Has a user-friendly interface built on top of the Visualisation ToolKit (VTK) library
• An extensible, modular architecture based on open standards
• Mainly used to analyse large datasets (petascale size) using distributed-memory computing resources
• Supports a wide range of data formats

Domain Specific Supported Data Types

ParaView Interface

Hands On Session 1:
ParaView Basics

Launch Paraview Bookmark

• Open your preferred browser (e.g. Chrome, Firefox) and goto: remotevis.pawsey.org.au

ParaView Window

Data

• Simulation of the flow of air around a heated and spinning disk

Open Data

/scratch/courses01/remote_vis/paraview/disk_out_ref.ex2

Load All Variables

• Click the checkbox to enable all variables in the properties panel and then click apply

Add Stream Tracer

• While having the data selected in the pipeline browser, add a stream tracer filter to the data

Add Stream Tracer cont.

• Set the seed type to "Point Source" and click apply to make the filter active

How to Search for Filters

• Click the filters menu in the top bar and select search

Add Tube Filter

• type all or part of "tube" to find the tube filter the press enter to add it

Activate Tube Filter

• Once again, you must click apply in the properties panel for the filter to activate
• You can also click the eye icon to apply

Add Glyph Filter

• Select the stream tracer object in the pipeline browser
• Add a Glyph filter from then filters bar

Configure Glyph Filter Properties

• Glyph type: Cone     Orientation: V
• Press the refresh scale factor button to automatically adjust scale
• You may need to manually adjust the scale and click apply to get good looking arrows
• Masking glyph mode: All points

Configure Glyph Colouring

• Scroll down to colouring and set the variable to use for colour to "temp"

Visualisation

• The final visualisation should look like this

Optional Experimentation

• Try experimenting with different representations for different data points

Hands on Session 2:
Paraview Client-Server Mode

Client-Server Mode

• To visualise large datasets, ParaView comes with the ability to run ParaView across multiple nodes using the distributed-memory paradigm
• We recommend to run the client on one node - make use of the graphics card on the corresponding node to visualise quickly at the front-end
• The ParaView servers run on other compute node(s) - The data communication is very fast, as the nodes are connected by an Infiniband cable (100Gb/s signaling rate)

Client-Server Mode

Launch Paraview Bookmark

• Open your preferred browser (e.g. Chrome, Firefox) and goto: https://remotevis.pawsey.org.au

Open the Server Configuration Window

• The server configuration for pawsey systems should already be loaded

Configure the Connection

• Check the training box if you are doing this in a live training session

Enter your Password

• The password prompt may appear in the console window you used to launch paraview

Wait for the Job to be accepted

• Be patient as this may take a while if you do not have a reservation

Confirm the Connection is Active

• If the pipeline browser shows the server name, then remote rendering is active

Optional Activities

Check the status of your Job

Open a terminal (MobaXterm if using Windows) on your laptop

ssh couxyz@zeus.pawsey.org.au

squeue -u $USER

Showing What Each Server Process is Rendering

For Efficient Parallel Rendering

Open "Settings" dialog box (Edit -> Settings)

For Efficient Parallel Rendering Cont.

"Render View" tab => "Remote/Parallel Rendering Options" => Set "Remote Render threshold" to 0 (Mbytes)

Hands on Session 2:
Large Data and Rendering Movies

Data

• Dataset - Dam break
• Column of water behind a membrane
• At t = 0s, membrane is removed, then:
• column of water collapses
• water impacts an obstacle
• creates a complicated flow structure
• 100 total timesteps

Open Data

/scratch/courses01/remote_vis/VTK/dbreak3d_vtutorial_..vtk

Activate Data in Render Window

• Click on the data in the pipeline broswer
• Click apply in the properties panel

Activate Data in Render Window cont.

• The data now shows in the render window

Colourise Data

• Set the colouring to be based on alpha.water in the properties panel

Generate Movie

• Click on the file menu and then save animation

Generate Movie Cont.

• Go to your user folder in /scratch where you have write permission
• Enter a file name and select type ".avi"

Generate Movie Cont.

• Select the advanced options cog in the top right

Generate Movie Cont.

• Choose a range of frames to render and click ok
• All 100 frames will take a long time so choose a range of around 10

Generate Movie Cont.

• The sofware will play though the anmiation as the frames are rendered and encoded

The final render (all 100 frames)

Copying the Video to Your Local Machine

Hands on Session 2:
Terminating ParaView Servers

Terminating ParaView Servers

Check running jobs for your user

squeue -u $USER

Cancel job by job_id

scancel job_id

Terminating ParaView Servers

Save and close the Paraview client

Terminating ParaView Servers

Close the terminal window opened by the server