Multi Cores Web based ParaView with all build-in filters

We are using ParaView at the moment to visualize OpenFoam files and we are planning to use Web based ParaView such as Trame or similar products (that sever to do heavy-lifting and browser rendering the results). We are looking for a multi-user, multi-core Paraview web app for openfoam parallel processing (using open mpi) where the files are hosted in AWS/cloud.

We ran some experiments with Trame using pvpython in ParaView-5.10.1-egl-MPI (running following command : …/ParaView-5.10.1-egl-MPI-Linux-Python3.9-x86_64/bin/pvpython -m paraview.apps.trame --venv .venv --trame-app pv_visualizer --data /efs --port 1234 --host

We could load OpenFoam files while it had three issues:

  1. It was very slow (No Multicores).
  2. No multi users
  3. Most of required and build-in ParaView filters and plugins were missing. We need to have filters in ParaView as listed below
  • Slice, Clip, Glyph, Resample, Resample with dataset, Calculator, streamline, streamline tracer, Integrate variables, python programmer, Extract block, Extract selection

We are running a feasibility study at this stage, before formally engaging to buy support service.

I was wondering

  1. if Web based ParaView can provide a robust and fast solution, potentially multi cores and multi users?

  2. if there is an in-house solution or product available to add all built-in filters/plugins to Web based ParaView?
    2.A If those products/filters available, how/where we could add them into Web Based ParaView?
    2.B. If those products/filters should be purchase from or developed by ParaView team, how we could engage your team? How many hours of support required?

Hi Aris,

Thanks for reaching out. Let me try to answer some of your questions.

  1. Multi-core: Are you talking about SMP filter or just using MPI to distribute the load? If it is mpi, you just need to use mpi with pvbatch.
  2. The multi-user tends to be specific on how you want to manage your deployment. One simple way to achieve that is to build the docker image and deploy it. Within that docker, you have the multi-user built-in. But that assume you are sharing the resources of 1 machine across several users. And based on your comment with MPI, that might not be desirable for your use case. Other options are possible like the reverse connection that we use when deploying trame onto a cluster.
  3. The missing filters are just artificial since the paraview-visualizer is still WIP and we just hard coded some value initially, but more could be added. That list is here.

So in summary:

  1. yes, it just need to be properly deployed and tuned
  2. Just need to extend the code so it could be extended by the user (a trivial thing to do)
    a. See here
    b. If you want us to expend that code base for you, just reach out for support. For that specific change and guidance for deployment, our eStore option could probably be enough.