Input Layout Question

DX9 removes unused elements from the signature. DX10 and up will keep them.

Thanks for the reply, but it doesn't really answer my question. My question is, what if I use this in the vertex shader:

struct VS_IN
{
float3 Pos : POSITION;
};

and use the input-layout based on:

struct Vertex
{
Vector3 Pos;
Vector3 Normal;
Vector3 UV;
};

Is this guaranteed to work in Direct3D 10+?

No, this will not work as the input signature is different.

Quote:

DX9 removes unused elements from the signature. DX10 and up will keep them.

OK, I found out though that you can have a fat vertex declaration, say:

D3D10_INPUT_ELEMENT_DESC layout[] ={    { L"POSITION", 0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 0,           D3D10_INPUT_PER_VERTEX_DATA, 0 },    { L"TEXTURE0", 0, DXGI_FORMAT_R32G32_FLOAT, 0, 12,           D3D10_INPUT_PER_VERTEX_DATA, 0 },    { L"NORMAL", 0, DXGI_FORMAT_R32G32B32_FLOAT, 1, 20,           D3D10_INPUT_PER_VERTEX_DATA, 0 },};

and a vertex shader input:

struct VS_IN{float3 Pos : POSITION;};

And you can create a valid input layout from the "fat" vertex description to the smaller input signature. So you can get from a larger vertex to a smaller, but you have to create a separate input layout for that mapping.

Quote:

No, this will not work as the input signature is different.

How come this works with the CascadedShadowMaps11 sample? They create one input layout for rendering both the scene and depth:

const D3D11_INPUT_ELEMENT_DESC layout_mesh[] =    {        { "POSITION",  0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 0,  D3D11_INPUT_PER_VERTEX_DATA, 0 },        { "NORMAL",    0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 12, D3D11_INPUT_PER_VERTEX_DATA, 0 },        { "TEXCOORD",  0, DXGI_FORMAT_R32G32_FLOAT,    0, 24, D3D11_INPUT_PER_VERTEX_DATA, 0 },    };        V_RETURN( pd3dDevice->CreateInputLayout(         layout_mesh, ARRAYSIZE( layout_mesh ),         m_pvsRenderSceneBlob[0]->GetBufferPointer(),        m_pvsRenderSceneBlob[0]->GetBufferSize(),         &m_pVertexLayoutMesh ) );

But in RenderCascadeShadow.hlsl they have:

struct VS_INPUT{    float4 vPosition    : POSITION;};

but they set pVertexLayoutMesh, which is based in the input signature in RenderCascadeScene.hlsl:

struct VS_INPUT{    float4 vPosition                        : POSITION;    float3 vNormal                          : NORMAL;    float2 vTexcoord                        : TEXCOORD0;};

I agree with Demirug.

D3D10 and above enhanced Vertext validation.

Each input-layout object is created based on a shader signature; this allows the API to validate the input-layout-object elements against the shader-input signature to make sure that there is an exact match of types and semantics. You can create a single input-layout object for many shaders, as long as all of the shader-input SIGNATURES EXACTLY MATCH.

In CascadedShadowMaps11 there are two places Input-Layout Objects are created.
one is for g_strUIEffectFile(line 628 to 639, DXUTgui.cpp), and the other is for RenderCascadeScene.hlsl(line 233 to 244, CascadedShadowsManager.cpp), while the vertex shader defined in RenderCascadeShadow.hlsl doesn't create primitives, so no D3D11_INPUT_ELEMENT_DESC struct and Input-Layout Object created for it.