bool
CLD3D11DeviceContext::Initialize(ID3D11DeviceContext *d3dDeviceContext) {
#if defined(OPENSUBDIV_HAS_DX11SDK) && \
(defined(OPENSUBDIV_HAS_CL_D3D11_H) || defined(OPENSUBDIV_HAS_CL_D3D11_EXT_H))
_d3dDeviceContext = d3dDeviceContext;
cl_int ciErrNum;
cl_platform_id cpPlatform = findPlatform();
ID3D11Device *device;
d3dDeviceContext->GetDevice(&device);
#if defined(OPENSUBDIV_HAS_CL_D3D11_H)
cl_context_properties props[] = {
CL_CONTEXT_D3D11_DEVICE_KHR, (cl_context_properties)device,
CL_CONTEXT_PLATFORM, (cl_context_properties)cpPlatform,
0
};
#elif defined(OPENSUBDIV_HAS_CL_D3D11_EXT_H)
cl_context_properties props[] = {
CL_CONTEXT_D3D11_DEVICE_NV, (cl_context_properties)device,
CL_CONTEXT_PLATFORM, (cl_context_properties)cpPlatform,
0
};
#endif
// get the number of GPU devices available to the platform
cl_uint numDevices = 0;
clGetDeviceIDs(cpPlatform, CL_DEVICE_TYPE_GPU, 0, NULL, &numDevices);
if (numDevices == 0) {
error("No CL GPU device found.\n");
return false;
}
// create the device list
cl_device_id *clDevices = new cl_device_id[numDevices];
clGetDeviceIDs(cpPlatform, CL_DEVICE_TYPE_GPU, numDevices, clDevices, NULL);
// we're cheating a little bit.
// try to find both cl_khr_d3d11_sharing and cl_nv_d3d11_sharing.
const char *extension = "_d3d11_sharing";
int clDeviceUsed = findExtensionSupportedDevice(clDevices, numDevices,
extension);
if (clDeviceUsed < 0) {
error("No device found that supports CL/D3D11 context sharing\n");
delete[] clDevices;
return false;
}
_clContext = clCreateContext(props, 1, &clDevices[clDeviceUsed],
NULL, NULL, &ciErrNum);
if (ciErrNum != CL_SUCCESS) {
error("Error %d in clCreateContext\n", ciErrNum);
delete[] clDevices;
return false;
}
_clCommandQueue = clCreateCommandQueue(_clContext, clDevices[clDeviceUsed],
0, &ciErrNum);
delete[] clDevices;
if (ciErrNum != CL_SUCCESS) {
error("Error %d in clCreateCommandQueue\n", ciErrNum);
return false;
}
return true;
#else
(void)d3dDeviceContext; // unused
return false;
#endif
}
// Initializes the OpenCL objects.
bool init_opencl(const int maxInputSize, const int maxOutputSize, const int maxWeightSize, const int maxBiasSize) {
cl_int status;
if (!setCwdToExeDir()) {
return false;
}
// Get the OpenCL platform.
platform = findPlatform("Altera");
if (platform == NULL) {
printf("ERROR: Unable to find Altera OpenCL platform.\n");
return false;
}
printf("Platform: %s\n", getPlatformName(platform).c_str());
// Query the available OpenCL devices.
scoped_array<cl_device_id> devices;
cl_uint num_devices;
devices.reset(getDevices(platform, CL_DEVICE_TYPE_ALL, &num_devices));
// We'll just use the first device.
device = devices[0];
printf("Device: %s\n", getDeviceName(device).c_str());
// Create the context.
context = clCreateContext(NULL, 1, &device, &oclContextCallback, NULL, &status);
checkError(status, "Failed to create context");
// Create the command queue.
queue = clCreateCommandQueue(context, device, CL_QUEUE_PROFILING_ENABLE, &status);
checkError(status, "Failed to create command queue");
// Create the program.
std::string binary_file = getBoardBinaryFile("waifu2x", device);
printf("Using AOCX: %s\n", binary_file.c_str());
program = createProgramFromBinary(context, binary_file.c_str(), &device, 1);
// Build the program that was just created.
status = clBuildProgram(program, 0, NULL, "", NULL, NULL);
checkError(status, "Failed to build program");
// Create the kernel - name passed in here must match kernel name in the
// original CL file, that was compiled into an AOCX file using the AOC tool
const char *kernel_name = "waifu2x"; // Kernel name, as defined in the CL file
kernel = clCreateKernel(program, kernel_name, &status);
checkError(status, "Failed to create kernel");
input_buf = clCreateBuffer(context, CL_MEM_READ_ONLY | CL_MEM_BANK_1_ALTERA,
maxInputSize * sizeof(float), NULL, &status);
checkError(status, "Failed to create buffer for input");
weight_buf = clCreateBuffer(context, CL_MEM_READ_ONLY | CL_MEM_BANK_2_ALTERA,
maxWeightSize * sizeof(float), NULL, &status);
checkError(status, "Failed to create buffer for weight");
bias_buf = clCreateBuffer(context, CL_MEM_READ_ONLY | CL_MEM_BANK_2_ALTERA,
maxBiasSize * sizeof(double), NULL, &status);
checkError(status, "Failed to create buffer for bias");
output_buf = clCreateBuffer(context, CL_MEM_WRITE_ONLY | CL_MEM_BANK_1_ALTERA,
maxOutputSize * sizeof(float), NULL, &status);
checkError(status, "Failed to create buffer for output");
return true;
}
bool
CLDeviceContext::Initialize() {
#ifdef OPENSUBDIV_HAS_CLEW
if (!clGetPlatformIDs) {
error("Error clGetPlatformIDs function not bound.\n");
return false;
}
#endif
cl_int ciErrNum;
cl_platform_id cpPlatform = findPlatform();
#if defined(_WIN32)
cl_context_properties props[] = {
CL_GL_CONTEXT_KHR, (cl_context_properties)wglGetCurrentContext(),
CL_WGL_HDC_KHR, (cl_context_properties)wglGetCurrentDC(),
CL_CONTEXT_PLATFORM, (cl_context_properties)cpPlatform,
0
};
#elif defined(__APPLE__)
CGLContextObj kCGLContext = CGLGetCurrentContext();
CGLShareGroupObj kCGLShareGroup = CGLGetShareGroup(kCGLContext);
cl_context_properties props[] = {
CL_CONTEXT_PROPERTY_USE_CGL_SHAREGROUP_APPLE, (cl_context_properties)kCGLShareGroup,
0
};
#else
cl_context_properties props[] = {
CL_GL_CONTEXT_KHR, (cl_context_properties)glXGetCurrentContext(),
CL_GLX_DISPLAY_KHR, (cl_context_properties)glXGetCurrentDisplay(),
CL_CONTEXT_PLATFORM, (cl_context_properties)cpPlatform,
0
};
#endif
#if defined(__APPLE__)
_clContext = clCreateContext(props, 0, NULL, clLogMessagesToStdoutAPPLE,
NULL, &ciErrNum);
if (ciErrNum != CL_SUCCESS) {
error("Error %d in clCreateContext\n", ciErrNum);
return false;
}
size_t devicesSize = 0;
clGetGLContextInfoAPPLE(_clContext, kCGLContext,
CL_CGL_DEVICES_FOR_SUPPORTED_VIRTUAL_SCREENS_APPLE,
0, NULL, &devicesSize);
int numDevices = int(devicesSize / sizeof(cl_device_id));
if (numDevices == 0) {
error("No sharable devices.\n");
return false;
}
cl_device_id *clDevices = new cl_device_id[numDevices];
clGetGLContextInfoAPPLE(_clContext, kCGLContext,
CL_CGL_DEVICES_FOR_SUPPORTED_VIRTUAL_SCREENS_APPLE,
numDevices * sizeof(cl_device_id), clDevices, NULL);
int clDeviceUsed = 0;
#else // not __APPLE__
// get the number of GPU devices available to the platform
cl_uint numDevices = 0;
clGetDeviceIDs(cpPlatform, CL_DEVICE_TYPE_GPU, 0, NULL, &numDevices);
if (numDevices == 0) {
error("No CL GPU device found.\n");
return false;
}
// create the device list
cl_device_id *clDevices = new cl_device_id[numDevices];
clGetDeviceIDs(cpPlatform, CL_DEVICE_TYPE_GPU, numDevices, clDevices, NULL);
const char *extension = "cl_khr_gl_sharing";
int clDeviceUsed = findExtensionSupportedDevice(clDevices, numDevices,
extension);
if (clDeviceUsed < 0) {
error("No device found that supports CL/GL context sharing\n");
delete[] clDevices;
return false;
}
_clContext = clCreateContext(props, 1, &clDevices[clDeviceUsed],
NULL, NULL, &ciErrNum);
#endif // not __APPLE__
if (ciErrNum != CL_SUCCESS) {
error("Error %d in clCreateContext\n", ciErrNum);
delete[] clDevices;
return false;
}
_clCommandQueue = clCreateCommandQueue(_clContext, clDevices[clDeviceUsed],
0, &ciErrNum);
delete[] clDevices;
if (ciErrNum != CL_SUCCESS) {
error("Error %d in clCreateCommandQueue\n", ciErrNum);
return false;
}
return true;
}
