Support vertex pulling #16826
Support vertex pulling #16826
Conversation
|
Please make sure to label your PR with "bug", "new feature" or "breaking change" label(s). |
|
Snapshot stored with reference name: Test environment: To test a playground add it to the URL, for example: https://snapshots-cvgtc2eugrd3cgfd.z01.azurefd.net/refs/pull/16826/merge/index.html#WGZLGJ#4600 Links to test babylon tools with this snapshot: https://playground.babylonjs.com/?snapshot=refs/pull/16826/merge To test the snapshot in the playground with a playground ID add it after the snapshot query string: https://playground.babylonjs.com/?snapshot=refs/pull/16826/merge#BCU1XR#0 |
|
Visualization tests for WebGPU |
|
WebGL2 visualization test reporter: |
|
It would be a great feature to have! However, I think we should strive to improve integration into the engine so that we/the user don't have to manually manage storage buffers and so that it remains as transparent as possible.
I wonder if we could just create the vertex buffers with STORAGE flags (in addition to the VERTEX + WRITE flags we currently use)... This way, the buffers should be usable in both pulling and non-pulling modes (to be tested, however, and I don't know if adding unnecessary flags has an impact on performance - STORAGE is not necessary in non-pulling mode, and VERTEX is not necessary in pulling mode). In this case, we could add a I'm probably missing a few things, but I think this would be a better way to support this mode. cc @sebavan for the discussion. |
|
Also, in your PG example, I think you would want to calculate the normal yourself by performing the vector product of two edges of the triangle, because the normal to a vertex is generally the average of the normals of the faces to which that vertex belongs, so it is not the true normal to the face. |
|
I really like where it is going :-) and I ll chat with Mike after his break. |
MiiBond
force-pushed
the
mbond/support-vertex-pulling
branch
from
a4c7ec9 to
8e92eed
Compare
|
WebGL2 visualization test reporter: |
|
Visualization tests for WebGPU |
|
I've updated the PR with:
I'm not sure if the way I did this is okay but it seems to work. |
|
Visualization tests for WebGPU |
|
WebGL2 visualization test reporter: |
|
Ah, looks like the index buffer in my test is being created as a Uint16Array and WGSL only supports u32 as a type... |
|
Okay, I've added support for 16-bit and 32-bit indices as well as non-indexed geometry. |
|
WebGL2 visualization test reporter: |
|
Visualization tests for WebGPU |
| @@ -1057,6 +1057,7 @@ export abstract class AbstractMesh extends TransformNode implements IDisposable, | |||
| protected _buildUniformLayout(): void { | |||
| this._uniformBuffer.addUniform("world", 16); | |||
| this._uniformBuffer.addUniform("visibility", 1); | |||
| this._uniformBuffer.addUniform("hasIndices", 1); | |||
There was a problem hiding this comment.
Shouldnot this force changing the shaders as well ?
There was a problem hiding this comment.
I don't think we need to add this uniform in the mesh ubo?
| @@ -2054,13 +2054,17 @@ export class Mesh extends AbstractMesh implements IGetSetVerticesData { | |||
|
|
|||
| const scene = this.getScene(); | |||
| const engine = scene.getEngine(); | |||
| const material = subMesh.getMaterial(); | |||
There was a problem hiding this comment.
I am not sure we should add this here as it is inside of the hot path. We should also get the info on the _draw callers ? cc @deltakosh
| // If there are vertex buffers that are not bound to the pipeline, AND there are storage buffers | ||
| // used by the shader with the same name, we will bind them to the current draw context because they | ||
| // are being used to do vertex pulling. | ||
| const availableVertexBuffers: { [key: string]: VertexBuffer } = (this._cacheRenderPipeline as any)._vertexBuffers; |
There was a problem hiding this comment.
Would be amazing to have @Popov72 feedback on it ?
There was a problem hiding this comment.
Let me take a look (as it is, I think it's probably too much work to do in the hot path).
|
Here's how I would do it: master...Popov72:Babylon.js:MiiBond-mbond/support-vertex-pulling Basically, I moved your code to To avoid running the code on every draw call and impacting performance, I exit early if the vertex pulling parameter (true/false) is the same as the previous call. This means we don't automatically handle a change in the vertex buffer list or if the index buffer changes after the first draw of the mesh. I don't think this is a problem, though, as it's unlikely that we'll need to add/remove a vertex buffer after displaying the mesh (we can still update a buffer). Let's see if anyone complains about this, and we'll take action if necessary. cc @sebavan |
Vertex pulling is where you do your own reading of vertex data in the vertex shader instead of relying on the standard vertex attribute pipeline. I'm not sure of the best approach in Babylon but I made some minimal changes to make it work. I'd appreciate some feedback and ideas for a better way to set this up.
First, some motivation:
I want vertex pulling as a way of selecting vertex data from neighboring vertices in the IBL Shadows voxelization shader. By retrieving normal info for the provoking vertex of a triangle, I can selectively swizzle the axes of the position to maximize rasterization area and avoid missing voxels. This eliminates the need for 3-pass voxelization and opens the door for doing realtime voxelization of animated geometry (in WebGPU only due to the need for 3D storage textures).
https://playground.babylonjs.com/?snapshot=refs/pull/16826/merge#XSNYAU#131
For vertex pulling to work, we typically need:
@builtin(vertex_index)to be sequential which is critical for fetching info about neighboring vertices.No changes are needed to Babylon for the first two requirements. For the 3rd point though, we want to do a
engine.drawArraysTypeeven though the mesh has indices. To do this, we setmaterial.useVertexPulling.The shader must also not declare vertex attributes.