Resolved Writing to RWTexture2DArray mipmaps in compute shader

Currently attempting to dive into compute shaders and trying to google some of this stuff is kicking my butt.

Pretty sure I've got the basics down but I hit a roadblock attempting to write to the mipmaps for a RWTexture2DArray I created. My current relevant code excerpts:

Code (CSharp):

1. RWTexture2DArray<float4> outputTexture : register(u0);
2.  
3. ...
4.  
5. [numthreads(32, 32, 1)]
6. void Convolve(uint3 ThreadID : SV_DispatchThreadID)
7. {
8.     ...
9.  
10.     outputTexture[ThreadID] = float4(1.0, 1.0, 1.0, 1.0);
11. }
12.  

It only seems to be able to write to mip 0 and outputTexture doesn't accept a float4 to specify a mip level.

I feel like this has to be a thing (this page certainly would seem to imply such a thing is possible: https://docs.microsoft.com/en-us/windows/win32/direct3d12/subresources) but I'm not exactly sure what to do with that information, especially in the context of Unity.

From what I gather, the subresource exists and I could sample it with a sampler, but how do I go about writing to it? Is there some Unity-specific thing I'm missing here? Unity's compute shader documentation is, uh... lacking, to put it nicely.


For some background context: the whole reason I'm using RWTexture2DArray explicitly is that there appears to be no way to write to a TextureCube so I'm doing the compute logic in a 6-deep RWTexture2DArray instead, then copying back to a cubemap RenderTexture afterwards. If there's a better way to do that, I'm all ears.

 

Is the destination mip level something that needs to be specified on the C# side? I haven't found any Unity documentation suggesting so which leads me to believe this needs to be done in the compute shader itself.

 

Bump.

 

Ah ha! Thank you, that's exactly what I was missing.