如何将计算着色器的结果传递给顶点着色器？(How to pass result from compute shader to vertex shader?)

 我的代码工作了。 首先，我创建一个缓冲区，通过CopyResource从计算着色器获取缓冲区。 
 
 然后我创建一个指针调用p1并从MappedResource1获取结果指针。 之后，我创建顶点缓冲区并渲染。 
 
 但我不认为这是从计算着色器获取顶点缓冲区的最佳方法： 
 
D3D11_BUFFER_DESC desc;
ZeroMemory(&desc, sizeof(desc));
mFirstBuffer->GetDesc(&desc);
desc.CPUAccessFlags = D3D11_CPU_ACCESS_READ;
desc.Usage = D3D11_USAGE_STAGING;
desc.BindFlags = 0;
desc.MiscFlags = 0;
if (SUCCEEDED(device->CreateBuffer(&desc, nullptr, &debugbuf)))
{
    context->CopyResource(debugbuf, mFirstBuffer.Get());
}
unsigned int stride = sizeof(InstanceData);
unsigned int offset = 0;
D3D11_MAPPED_SUBRESOURCE MappedResource1;
InstanceData *p1;

context->Map(debugbuf, 0, D3D11_MAP_READ, 0, &MappedResource1);
p1 = (InstanceData*)MappedResource1.pData;
context->Unmap(debugbuf, 0);

D3D11_SUBRESOURCE_DATA VertexBufferData;
VertexBufferData.pSysMem = p1;
VertexBufferData.SysMemPitch = 0;
VertexBufferData.SysMemSlicePitch = 0;
D3D11_BUFFER_DESC desc2;
desc2.ByteWidth = sizeof(InstanceData) * MillionParticleCount;
desc2.Usage = D3D11_USAGE_DEFAULT;
desc2.BindFlags = D3D11_BIND_VERTEX_BUFFER;
desc2.CPUAccessFlags = 0;
desc2.MiscFlags = 0;
desc2.StructureByteStride = 0;
device->CreateBuffer(&desc2, &VertexBufferData, &vertexBuffer);

context->IASetVertexBuffers(0, 1, &vertexBuffer, &stride, &offset);
 
 我已经阅读了很多样本​​，他们都只在一个文件中编写了计算着色器和顶点着色器，因此他们可以直接从计算着色器中获得结果。 
 
 在我的代码中，计算着色器和顶点着色器被写入不同的文件中。 我该怎么做才能优化我的代码？ 

My code worked. Firstly, I create a buffer to get buffer from compute shader by CopyResource.
 
Then I create a pointer call p1 and get result pointer from MappedResource1. After that, I create vertex buffer and render.
 
But I don't think it's the best way to get vertex buffer from compute shader:
 
D3D11_BUFFER_DESC desc;
ZeroMemory(&desc, sizeof(desc));
mFirstBuffer->GetDesc(&desc);
desc.CPUAccessFlags = D3D11_CPU_ACCESS_READ;
desc.Usage = D3D11_USAGE_STAGING;
desc.BindFlags = 0;
desc.MiscFlags = 0;
if (SUCCEEDED(device->CreateBuffer(&desc, nullptr, &debugbuf)))
{
    context->CopyResource(debugbuf, mFirstBuffer.Get());
}
unsigned int stride = sizeof(InstanceData);
unsigned int offset = 0;
D3D11_MAPPED_SUBRESOURCE MappedResource1;
InstanceData *p1;

context->Map(debugbuf, 0, D3D11_MAP_READ, 0, &MappedResource1);
p1 = (InstanceData*)MappedResource1.pData;
context->Unmap(debugbuf, 0);

D3D11_SUBRESOURCE_DATA VertexBufferData;
VertexBufferData.pSysMem = p1;
VertexBufferData.SysMemPitch = 0;
VertexBufferData.SysMemSlicePitch = 0;
D3D11_BUFFER_DESC desc2;
desc2.ByteWidth = sizeof(InstanceData) * MillionParticleCount;
desc2.Usage = D3D11_USAGE_DEFAULT;
desc2.BindFlags = D3D11_BIND_VERTEX_BUFFER;
desc2.CPUAccessFlags = 0;
desc2.MiscFlags = 0;
desc2.StructureByteStride = 0;
device->CreateBuffer(&desc2, &VertexBufferData, &vertexBuffer);

context->IASetVertexBuffers(0, 1, &vertexBuffer, &stride, &offset);
 
I have read many sample, all of them write compute shader and vertex shader together in only one file and therefor they can get result from compute shader directly.
 
In my code, compute shader and vertex shader are write in different files. How should I do to optimize my code?

原文：https://stackoverflow.com/questions/37053051