ParaView

ParaView is a powerful open-source data analysis and visualization tool that has been containerized for easy deployment. Equipped with advanced features such as NVIDIA IndeX and OptiX, it enables efficient analysis and visualization of large and complex datasets by utilizing parallel processing and distributed computing techniques. NVIDIA OptiX gives ParaView the power to render physically accurate lighting and material, enabling users to create high-quality visualizations, while the integration of NVIDIA IndeX allows for improved data indexing and querying, further enhancing the analysis capabilities of the application.

System requirements

Before running the NGC ParaView container please ensure your system meets the following requirements.

• One of the following container runtimes
  • nvidia-docker
  • Singularity >= 3.1
• One of the following NVIDIA GPU(s)
  • Pascal(sm60)
  • Volta (sm70)
  • Ampere (sm80)
  • Hopper (sm90)

x86_64

• One of the following CUDA driver versions
  • >= r450 (>=.80.02)

Examples

The following examples demonstrate using the NGC ParaView container to render wavelet source using the latest ParaView image.

export IMAGE_TAG=nvcr.io/nvidia-hpcvis/paraview:egl-py3-5.11.0

Running ParaView Server

To start the ParaView Server on one node with GPUs enabled, you can use the following command in the terminal:

docker run --gpus all -p 11111:11111 $IMAGE_TAG ./usr/local/paraview/bin/pvserver

To connect to the ParaView server using the command line, use the following command:

./bin/paraview --server-url=cs://HOST:11111

Note: Load all the plugins you plan to use on your local client before connecting to a server.

Using the OptiX Pathtracing Backend on ParaView Server

To enable the NVIDIA OptiX pathtracing backend and test it with some synthetic data, follow these steps:

1. In the Properties window (lower-left by default, enabled in the View menu if it is missing), go to the View section > Ray Traced Rendering subsection and check Enable Ray Tracing.
2. In the Ray Traced Rendering subsection, go to the Backend drop down menu and select OptiX pathtracer.
3. Go to the Sources menu > Data Objects > Wavelet, then Properties > Apply.
4. Right-click the Wavelet source and select Add Filter > Common > Contour, then Properties > Apply.
5. You should now see a ray traced Wavelet source with proper illumination and shadows. Try enabling/disabling ray tracing a few times for comparison.

Running ParaViewWeb

The Visualizer Web application allows for a ParaView-like experience within a web browser. This is made possible through the ParaViewWeb library, which contains all the necessary components for building the user interface and facilitating data access via WebSocket connectivity to a ParaView server. The Visualizer application connects all of these components in a meaningful way.

docker run --gpus all -p 8080:8080 -v /host_data_dir:/data $IMAGE_TAG ./usr/local/paraview/bin/pvpython /usr/local/paraview/share/paraview-5.11/web/visualizer/server/pvw-visualizer.py --content /usr/local/paraview/share/paraview-5.11/web/visualizer/www --port 8080 --data /data -i 0.0.0.0

Note: If you aren't using the latest ParaView image, be sure to update the directories in the command above.

Using the NVIDIA IndeX plugin on ParaViewWeb

To verify that IndeX is functioning properly, follow these steps to render a generated volume using the NVIDIA IndeX renderer:

1. Create a Wavelet by clicking on the + sign and selecting Wavelet.
2. Select the scalar(p1)RTData option by clicking on the droplet dropdown menu and choosing it (be sure to deselect Solid color).
3. Select the NVIDIA IndeX renderer by clicking on the eye dropdown menu and selecting it.
4. Once NVIDIA IndeX initializes, the volume should be visible.

FAQ

• Q: When creating a Wavelet, ParaView crashes with the following error:

vtkSIProxy.cxx:422    ERR| vtkSISourceProxy (0x2296bb70): Failed to create 'vtknvindex_representation'. This typically means that ParaView does not know about the request class to create an instance of if. Ensure that it has been correctly wrapped using the client-server wrappers and the wrapping has been initialized. Note class names are case-sensitive. Check for typos. Aborting for debugging purposes.

If you're receiving this message, it means that the NVIDIA IndeX plugin has been loaded on the remote server but not locally. To resolve this issue, be sure to load the plugin on both the remote and local sides as specified above.

Suggested Reading

ParaView Superbuild Repository

Download ParaView

ParaView User's Guide