Pass data to fragment shader or call fragment shader

[davids91]

Hi and thank you for this library, seems like an amazing addition to the ecosystem!
Is there a way to make the calculated data available to a fragment shader to display the results?

For my specific purposes I would need to avoid copying to the CPU as the computed data is too much to copy every frame.

[davids91]

Hey! I see no fragment shaders in dsenyushkindev@27d0148

Did I miss something?

[Kjolnyr]

Hi @davids91 ,

The crate has been dead for the last 2 years, I've finally decided to resurrect it. While doing so I closed all the issues since they were all really old so I wanted to have a fresh (re)start.

I'll leave this one open then since you still request the feature, and I'll try to work on that ASAP.

Thank again for the kinds words!

[davids91]

I since then moved on to implement most of what I needed in raw wgpu; But Nonetheless I still think this would be a great addition to the library! Feel free to close it though if you wouldn't want to have this feature!

[sesteel]

I also want this.

[BrianWiz]

If I understand you correctly, I think I need something similar. In my case I need to have a storage buffer that a compute shader writes to, and a material that reads from it in its vertex shader by assigning the handle to a uniform on the material. In that case, all I need is the Handle<ShaderStorageBuffer> to the storage that was added to the compute worker. And what you end up with is no need for copying back to the CPU at all, it just stays on the GPU. I created a ticket for this specific case: #21

[davids91]

Yes, from experience this works like that!
See my implementation within the bevy pipeline:
https://github.com/Ministry-of-Voxel-Affairs/VoxelHex/blob/8c51f2de1d85a37fd22f995198123dfe8f5987ef/src/raytracing/bevy/pipeline/mod.rs#L127

Except the compute shader doesn't write to a uniformbuffer, but a texture:
https://github.com/Ministry-of-Voxel-Affairs/VoxelHex/blob/8c51f2de1d85a37fd22f995198123dfe8f5987ef/src/raytracing/bevy/viewport_render.wgsl#L804

Furthermore synchronization is handled within WGPU so data should be guarded from the race conditions by the scheduler.