Cécile Betmont
Solution Engineer
HPC platform
Launch compute tasks in a few lines of code or a few clicks on Tasq, our HPC platform.

Bella Render on Qarnot Cloud - Documentation

January 18, 2024 - 3D, Documentation


Bella is a spectral render engine which models the complexities of optical physics, using analogies grounded in the real world, to produce a system that behaves in a predictable and intuitive way. Running your Bella engine on Qarnot is as easy as uploading your case and launching a script. Here’s a walkthrough of the steps.



The test use case uses Bella v.23.6.0.

Release year     Version
jan 202423.6.0


If you are interested in another version, please send us an email at qlab@qarnot.com.



Before launching the case, please ensure that the following prerequisites have been met.


Test cases

This use case is based on the 3D Bella render tutorial simulation. The documentation is available here : bella_scenes_documentation. Some use cases are available on the bella render documentation. We suggest here to compute a simple orange-juice scene.
You can download the input scene here : orange-juice.bsz and drag it into a folder named ‘input’ to launch the calculation.


Launching a batch case


Copy the following code in a Python script and save it next to the input folder you unzipped before. Be sure you have copied your authentication token in the script (instead of <<<MY_SECRET_TOKEN>>>) to be able to launch the task on Qarnot.




At any given time, you can monitor the status of your task on Tasq :


Bella status on TasQ



Once the task is deployed, it should take around 15 minutes to run. You should also now have a result folder in the output bucket and on your computer containing all the numerical results. You can open the out_orange_0_0000.png (~40Mb) and out_orange_0_0001.png (~40Mb) files to view the results.


The expected results are as follow :




You could test in the same way any other scenes available in the tutorial.


Available parameters :


  * RESSOURCE_PATH : Default /job
  * BENCHMARK : Activate the benchmark options (True of False)
  * RESOLUTION : Override the resolution in the bsx file. (eg : 60x60)
  * RENDER_TURN_TABLE : Tells the CLI to render a turntable animation of the input file. (True or False)
  * TURN_TABLE_ANGLE : Sets the angle between frames for a turntable animation (eg : 160)
  * TURN_TABLE_STEP : Sets the total angle covered by a turntable animation. (eg : 40)
  * PARSE_FRAGMENTS : Provide a BSA fragment to be parsed by the scene (eg : camera.resolution=vec2(100 100); beautyPass.targetNoise=10u; )
  * FINAL_BSI_DIR : The relative path where bsi objects will be saved (saved as according to beautyPass:saveBsi). It will be available in the output bucket.
  * FINAL_BSI_NAME : The name of bsi objects will be saved
  * OUTPUT_PATH : The output relative path where outputs will be saved. It will be available in the output bucket with name <output_path>_0_xxxx with xxxx the image increment number (eg : out_orange_0_0001.png).
  * LICENSE_PATH : The path to the license in the input bucket.



The exemple below is a simple use case working without license. 
Of course, you can bring your own license by uploading the license file in the input bucket and setting its path in the LICENSE_PATH parameter.


Wrapping up

That’s it! If you have any questions, please contact qlab@qarnot.com and we will help you with pleasure!


Share on networks