Hi,
DirectX allows me to bind an array of render target views, or a render target view created from an array resource, and select which to write to using the SV_RenderTargetArrayIndex semantic output from a geometry shader. All good. But as far as I can see (from MSDN) there's a limitation in that all bound render targets have to share the same depth buffer.
The thing I'm working on would require each target to have its own associated depth buffer, essentially writing each triangle to eight different targets, each with different transforms but otherwise identical render states etc.
Binding a depth stencil view created from a texture2darray resource doesn't seem to work. There's no obvious failure on creating the resource, but nothing gets rendered when I try to use such a resource. I guess my semantic pedantic argument that this is should count as a single view doesn't hold weight :-)
I've already thought of a slightly hacky workaround, but does anyone know of a way to do what I'm trying without jumping through too many hoops?
Thanks
--
