Replies: 2 comments
-
Hello @WeirodoK, Here are some high-level answers to give an intuition. I would refer to the Sionna RT Technical report for more details.
Generally, ray tracing cost increases with the number of primitives (triangles, i.e. geometric complexity) in the scene.
The computation cost grows linearly with the number of paths. But that's also the dimension that is massively parallelized on the GPU, so you should find that many, many paths can be handled in parallel without a great increase in runtime, until the GPU is saturated.
I don't think this would influence the run time much. |
Beta Was this translation helpful? Give feedback.
-
This is not an issue with the Sionna code |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
I am currently working on optimizing a scene in Sionna, and I would like to understand the time complexity with respect to the following parameters:
Number of Objects in the Scene: As the number of materials (reflective surfaces, objects) increases in the scene, how does this affect the computational complexity? Specifically, does the complexity grow linearly or quadratically with the number of objects?
Number of Scattered Paths: When the number of scattered paths increases (due to path tracing or more detailed ray modeling), how does the time complexity change? Is it a linear or higher-order growth?
Number of Parameters per Material to Infer: How does the time complexity change when more parameters need to be inferred per material (e.g., conductivity, relative permittivity)?
A more theoretical explanation of how these factors contribute to the overall time complexity in big-O notation would be greatly appreciated. Additionally, if there are any related documents or reference materials available, could you kindly share them for further reading?
Beta Was this translation helpful? Give feedback.
All reactions