Sign in to follow this  
jakovo

Using some ConstantBuffer values causing CreateShader to fail

Recommended Posts

I think I still need to understand better how to use constant buffers.

 

I have a constant buffer with just two floats declared:

 

C++ struct:

struct ConstantBuffer
{
    float value0;
    float value1;
} g_CBData;

.

 

And in the Vertex Shader I'm just trying to multiply the color by the Sin() of any of these two values:

cbuffer cbPerFrame : register (b0)
{
    float  value0;
    float  value1;
};


struct VSInput
{
    float3 Pos : POSITION;
    float4 Col : COLOR;
};


struct VSOutput
{
    float4 Pos : SV_POSITION;
    float4 Col : COLOR;
};


VSOutput main( VSInput vs_in )
{
    VSOutput vs_out;

    vs_out.Pos = float4( vs_in.Pos, 1 );
    vs_out.Col = vs_in.Col * sin( value1 ); // "value0" runs ok, "value1" causes E_INVALIDARG

    return vs_out;
}

 

If I use "value0" everything runs well, however, if I use "value1" then I get an E_INVALIDARG when creating the VertexShader:

R = g_pDirect3D->CreateVertexShader( pVSData, vsSize, nullptr, &g_pVertexShader );

.

 

The weird thing is that if I don't use the sin() function (i.e. I just multiply Color * value1):

vs_out.Col = vs_in.Col * value0; // runs ok
vs_out.Col = vs_in.Col * value1; // runs ok
.

 

Then the values are read correctly.

 

Also I can't multiply value0 * value1, nor sum or anything between these two values.. when I do, I just get the E_INVALIDARG error.

 

Is there any strange rule when using the values from a constant buffer?

 

Thanks

 

 

Share this post


Link to post
Share on other sites

First of all, instead of blindfoldetly quessing whats the problem, you should always check what the actual error is. There is a nice method for this, its not like DX won't give you any information whats wrong...

ID3DBlob* pErrorBlob = nullptr;
ID3DBlob* pVertexBlob = nullptr;

if(FAILED(D3DX11CompileFromFile(lpShaderName, nullptr, nullptr, "mainVS", "vs_5_0", 0, 0, nullptr, &pVertexBlob, &pErrorBlob, nullptr)))
{
    if(pErrorBlob)
    {
        char* compileErrors = (char*)(pErrorBlob->GetBufferPointer());
        sys::log->Out(compileErrors);
        pErrorBlob->Release();
    }
    else
        sys::log->Out("Effect file not found.");

    throw d3dException();
}

Try to implement this, and somehow output the compileError-string in readable manner. What does it tell you? Post the message here if you still don't know whats going on, I'll have a look at it.

 

If I had to quess, I'd it crashes when you use value1, but not value0. I can't tell for sure, but from what I get it always crashes when only value1 is used, right? If no variable out of cbuffer is used, this buffer isn't even compiled, and it could be so too that if you don't use value0, it won't get accounted for and causes some sort of alignement/whatever problem. Well, I'm waiting for the error message biggrin.png

Edited by Juliean

Share this post


Link to post
Share on other sites

Thanks Juliean,

 

In this particular case I'm precompiling the shader (I'm targeting Win8, so no effects framework, etc), and then I'm loading the compiled data to my app.

 

However, the error is not at compile time (it compiles just fine), but at creation time when calling:

g_pDirect3D->CreateVertexShader( pVSData, vsSize, nullptr, &g_pVertexShader );

So it just doesn't make sense to me that by using "sin( value0 )"  the shader compiles, creates and runs without problems, but when using "sin( value1 )" the shader still compiles fine but cannot be created.

 

I just don't know what might be causing it unsure.png

 

 

thanks!

Share this post


Link to post
Share on other sites

In the end the problem was that I wasn't loading the compiled shader as a binary file, so, sometimes the file read was interrupted before the actual end, thus the CreateVertexShader() caused an E_INVALIDARG because the size of the vs data did not match the actual size of the shader.

 

dogh!...

 

anyways!... thanks for the help!

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this