CUDA and Direct3D consistency

CUDA And Direct3D Uniformity

Direct3D 9Ex、Direct3D 10 and Direct3D 11 Support Direct3D Interoperability .

CUDA Context can only be used with Direct3D Device interoperability ： Direct3D 9Ex The device must be set to D3DDEVTYPE_HAL Of DeviceType And use D3DCREATE_HARDWARE_VERTEXPROCESSING logo BehaviorFlags establish ; Direct3D 10 and Direct3D 11 The device must be in DriverType Set to D3D_DRIVER_TYPE_HARDWARE Created in case of .

Can be mapped to CUDA Address space Direct3D Resources are Direct3D buffer 、 Texture and surface . These resources use cudaGraphicsD3D9RegisterResource()、cudaGraphicsD3D10RegisterResource() and cudaGraphicsD3D11RegisterResource() register .

The following code example uses the kernel to dynamically modify 2D width x height grid .

Direct3D 9 Version:

IDirect3D9* D3D;
IDirect3DDevice9* device;
struct CUSTOMVERTEX {
    FLOAT x, y, z;
    DWORD color;
};
IDirect3DVertexBuffer9* positionsVB;
struct cudaGraphicsResource* positionsVB_CUDA;

int main()
{
    int dev;
    // Initialize Direct3D
    D3D = Direct3DCreate9Ex(D3D_SDK_VERSION);

    // Get a CUDA-enabled adapter
    unsigned int adapter = 0;
    for (; adapter < g_pD3D->GetAdapterCount(); adapter++) {
        D3DADAPTER_IDENTIFIER9 adapterId;
        g_pD3D->GetAdapterIdentifier(adapter, 0, &adapterId);
        if (cudaD3D9GetDevice(&dev, adapterId.DeviceName)
            == cudaSuccess)
            break;
    }

     // Create device
    ...
    D3D->CreateDeviceEx(adapter, D3DDEVTYPE_HAL, hWnd,
                        D3DCREATE_HARDWARE_VERTEXPROCESSING,
                        &params, NULL, &device);

    // Use the same device
    cudaSetDevice(dev);

    // Create vertex buffer and register it with CUDA
    unsigned int size = width * height * sizeof(CUSTOMVERTEX);
    device->CreateVertexBuffer(size, 0, D3DFVF_CUSTOMVERTEX,
                               D3DPOOL_DEFAULT, &positionsVB, 0);
    cudaGraphicsD3D9RegisterResource(&positionsVB_CUDA,
                                     positionsVB,
                                     cudaGraphicsRegisterFlagsNone);
    cudaGraphicsResourceSetMapFlags(positionsVB_CUDA,
                                    cudaGraphicsMapFlagsWriteDiscard);

    // Launch rendering loop
    while (...) {
        ...
        Render();
        ...
    }
    ...
}
void Render()
{
    // Map vertex buffer for writing from CUDA
    float4* positions;
    cudaGraphicsMapResources(1, &positionsVB_CUDA, 0);
    size_t num_bytes; 
    cudaGraphicsResourceGetMappedPointer((void**)&positions,
                                         &num_bytes,  
                                         positionsVB_CUDA));

    // Execute kernel
    dim3 dimBlock(16, 16, 1);
    dim3 dimGrid(width / dimBlock.x, height / dimBlock.y, 1);
    createVertices<<<dimGrid, dimBlock>>>(positions, time,
                                          width, height);

    // Unmap vertex buffer
    cudaGraphicsUnmapResources(1, &positionsVB_CUDA, 0);

    // Draw and present
    ...
}

void releaseVB()
{
    cudaGraphicsUnregisterResource(positionsVB_CUDA);
    positionsVB->Release();
}

__global__ void createVertices(float4* positions, float time,
                               unsigned int width, unsigned int height)
{
    unsigned int x = blockIdx.x * blockDim.x + threadIdx.x;
    unsigned int y = blockIdx.y * blockDim.y + threadIdx.y;

    // Calculate uv coordinates
    float u = x / (float)width;
    float v = y / (float)height;
    u = u * 2.0f - 1.0f;
    v = v * 2.0f - 1.0f;

    // Calculate simple sine wave pattern
    float freq = 4.0f;
    float w = sinf(u * freq + time)
            * cosf(v * freq + time) * 0.5f;

    // Write positions
    positions[y * width + x] =
                make_float4(u, w, v, __int_as_float(0xff00ff00));
}

Direct3D 10 Version

ID3D10Device* device;
struct CUSTOMVERTEX {
    FLOAT x, y, z;
    DWORD color;
};
ID3D10Buffer* positionsVB;
struct cudaGraphicsResource* positionsVB_CUDA;
            
int main()
{
    int dev;
    // Get a CUDA-enabled adapter
    IDXGIFactory* factory;
    CreateDXGIFactory(__uuidof(IDXGIFactory), (void**)&factory);
    IDXGIAdapter* adapter = 0;
    for (unsigned int i = 0; !adapter; ++i) {
        if (FAILED(factory->EnumAdapters(i, &adapter))
            break;
        if (cudaD3D10GetDevice(&dev, adapter) == cudaSuccess)
            break;
        adapter->Release();
    }
    factory->Release();
            
    // Create swap chain and device
    ...
    D3D10CreateDeviceAndSwapChain(adapter, 
                                  D3D10_DRIVER_TYPE_HARDWARE, 0, 
                                  D3D10_CREATE_DEVICE_DEBUG,
                                  D3D10_SDK_VERSION, 
                                  &swapChainDesc, &swapChain,
                                  &device);
    adapter->Release();

    // Use the same device
    cudaSetDevice(dev);

    // Create vertex buffer and register it with CUDA
    unsigned int size = width * height * sizeof(CUSTOMVERTEX);
    D3D10_BUFFER_DESC bufferDesc;
    bufferDesc.Usage          = D3D10_USAGE_DEFAULT;
    bufferDesc.ByteWidth      = size;
    bufferDesc.BindFlags      = D3D10_BIND_VERTEX_BUFFER;
    bufferDesc.CPUAccessFlags = 0;
    bufferDesc.MiscFlags      = 0;
    device->CreateBuffer(&bufferDesc, 0, &positionsVB);
    cudaGraphicsD3D10RegisterResource(&positionsVB_CUDA,
                                      positionsVB,
                                      cudaGraphicsRegisterFlagsNone);
                                      cudaGraphicsResourceSetMapFlags(positionsVB_CUDA,
                                      cudaGraphicsMapFlagsWriteDiscard);

    // Launch rendering loop
    while (...) {
        ...
        Render();
        ...
    }
    ...
}
void Render()
{
    // Map vertex buffer for writing from CUDA
    float4* positions;
    cudaGraphicsMapResources(1, &positionsVB_CUDA, 0);
    size_t num_bytes; 
    cudaGraphicsResourceGetMappedPointer((void**)&positions,
                                         &num_bytes,  
                                         positionsVB_CUDA));

    // Execute kernel
    dim3 dimBlock(16, 16, 1);
    dim3 dimGrid(width / dimBlock.x, height / dimBlock.y, 1);
    createVertices<<<dimGrid, dimBlock>>>(positions, time,
                                          width, height);

    // Unmap vertex buffer
    cudaGraphicsUnmapResources(1, &positionsVB_CUDA, 0);

    // Draw and present
    ...
}

void releaseVB()
{
    cudaGraphicsUnregisterResource(positionsVB_CUDA);
    positionsVB->Release();
}

__global__ void createVertices(float4* positions, float time,
                               unsigned int width, unsigned int height)
{
    unsigned int x = blockIdx.x * blockDim.x + threadIdx.x;
    unsigned int y = blockIdx.y * blockDim.y + threadIdx.y;

    // Calculate uv coordinates
    float u = x / (float)width;
    float v = y / (float)height;
    u = u * 2.0f - 1.0f;
    v = v * 2.0f - 1.0f;

    // Calculate simple sine wave pattern
    float freq = 4.0f;
    float w = sinf(u * freq + time)
            * cosf(v * freq + time) * 0.5f;
            
    // Write positions
    positions[y * width + x] =
                make_float4(u, w, v, __int_as_float(0xff00ff00));
}

Direct3D 11 Version

ID3D11Device* device;
struct CUSTOMVERTEX {
    FLOAT x, y, z;
    DWORD color;
};
ID3D11Buffer* positionsVB;
struct cudaGraphicsResource* positionsVB_CUDA;

int main()
{
    int dev;
    // Get a CUDA-enabled adapter
    IDXGIFactory* factory;
    CreateDXGIFactory(__uuidof(IDXGIFactory), (void**)&factory);
    IDXGIAdapter* adapter = 0;
    for (unsigned int i = 0; !adapter; ++i) {
        if (FAILED(factory->EnumAdapters(i, &adapter))
            break;
        if (cudaD3D11GetDevice(&dev, adapter) == cudaSuccess)
            break;
        adapter->Release();
    }
    factory->Release();

    // Create swap chain and device
    ...
    sFnPtr_D3D11CreateDeviceAndSwapChain(adapter, 
                                         D3D11_DRIVER_TYPE_HARDWARE,
                                         0, 
                                         D3D11_CREATE_DEVICE_DEBUG,
                                         featureLevels, 3,
                                         D3D11_SDK_VERSION, 
                                         &swapChainDesc, &swapChain,
                                         &device,
                                         &featureLevel,
                                         &deviceContext);
    adapter->Release();

    // Use the same device
    cudaSetDevice(dev);

    // Create vertex buffer and register it with CUDA
    unsigned int size = width * height * sizeof(CUSTOMVERTEX);
    D3D11_BUFFER_DESC bufferDesc;
    bufferDesc.Usage          = D3D11_USAGE_DEFAULT;
    bufferDesc.ByteWidth      = size;
    bufferDesc.BindFlags      = D3D11_BIND_VERTEX_BUFFER;
    bufferDesc.CPUAccessFlags = 0;
    bufferDesc.MiscFlags      = 0;
    device->CreateBuffer(&bufferDesc, 0, &positionsVB);
    cudaGraphicsD3D11RegisterResource(&positionsVB_CUDA,
                                      positionsVB,
                                      cudaGraphicsRegisterFlagsNone);
    cudaGraphicsResourceSetMapFlags(positionsVB_CUDA,
                                    cudaGraphicsMapFlagsWriteDiscard);

    // Launch rendering loop
    while (...) {
        ...
        Render();
        ...
    }
    ...
}
void Render()
{
    // Map vertex buffer for writing from CUDA
    float4* positions;
    cudaGraphicsMapResources(1, &positionsVB_CUDA, 0);
    size_t num_bytes; 
    cudaGraphicsResourceGetMappedPointer((void**)&positions,
                                         &num_bytes,  
                                         positionsVB_CUDA));

    // Execute kernel
    dim3 dimBlock(16, 16, 1);
    dim3 dimGrid(width / dimBlock.x, height / dimBlock.y, 1);
    createVertices<<<dimGrid, dimBlock>>>(positions, time,
                                          width, height);

    // Unmap vertex buffer
    cudaGraphicsUnmapResources(1, &positionsVB_CUDA, 0);

    // Draw and present
    ...
}

void releaseVB()
{
    cudaGraphicsUnregisterResource(positionsVB_CUDA);
    positionsVB->Release();
}

    __global__ void createVertices(float4* positions, float time,
                          unsigned int width, unsigned int height)
{
    unsigned int x = blockIdx.x * blockDim.x + threadIdx.x;
    unsigned int y = blockIdx.y * blockDim.y + threadIdx.y;

// Calculate uv coordinates
    float u = x / (float)width;
    float v = y / (float)height;
    u = u * 2.0f - 1.0f;
    v = v * 2.0f - 1.0f;

    // Calculate simple sine wave pattern
    float freq = 4.0f;
    float w = sinf(u * freq + time)
            * cosf(v * freq + time) * 0.5f;

    // Write positions
    positions[y * width + x] =
                make_float4(u, w, v, __int_as_float(0xff00ff00));
}

SLI Uniformity

In having more than one GPU In the system , All support CUDA Of GPU Both can pass CUDA The driver and runtime are accessed as separate devices . However , When the system is in SLI Mode time , There are special precautions as follows .

First , In a GPU On the one CUDA Allocation in the device will consume other GPU In the memory , these GPU yes Direct3D or OpenGL The equipment SLI Part of the configuration . therefore , The allocation may fail earlier than expected .

secondly , The application should create multiple CUDA Context , One for the SLI Each of the GPU. Although this is not a strict requirement , But it avoids unnecessary data transmission between devices . The application can cudaD3D[9|10|11]GetDevices() be used for Direct3D and cudaGLGetDevices() be used for OpenGL call , To identify the device currently performing rendering CUDA Device handle and next frame . In view of this information , Applications usually choose the right device and put Direct3D or OpenGL Resources are mapped to by cudaD3D[9|10|11]GetDevices() Or when deviceList Parameter set to cudaD3D[9|10 |11]DeviceListCurrentFrame or cudaGLDeviceListCurrentFrame.

Please note that , from cudaGraphicsD9D[9|10|11]RegisterResource and cudaGraphicsGLRegister[Buffer|Image] The returned resources can only be used on the device where the registration occurs . therefore , stay SLI Configuration in progress , When in different CUDA When calculating data of different frames on the device , It is necessary to register resources for each device separately .

of CUDA How to separate the runtime from Direct3D and OpenGL Details of interoperability , see also Direct3D Interoperability and OpenGL Interoperability .