clusterizer: Implement experimental triangle mask generator #553
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This change computes triangle masks for each frustum slice, following "GPU-Driven Rendering Pipelines" from SIGGRAPH 2015. The resulting masks can be used to quickly cull individual triangles, which only requires classifying the camera position in cube frustum face in the object space and selecting the appropriate set of triangle bits. The masks can also be used to reject the entire clusters, or entire triangle groups, by comparing all of them to 0 in task shader and/or comparing each 32-triangle group to 0 in mesh shader before emitting that triangle group.
…asks This provides a more uniform interface wrt meshopt_computeClusterBounds and is convenient when the data isn't following the meshlet format (eg consecutive chunks of index buffer are designated as clusters).
…neric We now reuse less computation and as such masks take ~5% more time to compute, but this code is a little harder to get wrong and will be easier to extend to 24 regions. Also adjust demo code to just perform classification on one side.
Also remove the main nested loop as a single loop is less error prone and ~as efficient, and add a C++ template helper. This change has been validated end-to-end on niagara renderer, and results in ~10-15% triangles culled on average.
6 and 8 don't seem to be quite good enough; 24 is fairly reasonable. This interface is very rigid so we will probably need to change this so that meshopt only implements single-region classification.
|
Per https://zeux.io/2023/04/28/triangle-backface-culling/ the results here are not that interesting so I'll leave this up for a bit but it looks unlikely that this will be merged. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This change computes triangle masks for each frustum slice, following "GPU-Driven Rendering Pipelines" from SIGGRAPH 2015.
The resulting masks can be used to quickly cull individual triangles, which only requires classifying the camera position in cube frustum face in the object space and selecting the appropriate set of triangle bits.
The masks can also be used to reject the entire clusters, or entire triangle groups, by comparing all of them to 0 in task shader and/or comparing each 32-triangle group to 0 in mesh shader before emitting that triangle group.
Note: this change is not fully tested, so it's not ready (may contain errors in sign/winding/etc.; also needs demo code cleanup)