/**
* Prints the device properties.
* @param file the file to print out the properties.
*/
void printDevProp(FILE* file) {
cudaDeviceProp devProp = getDeviceProperties();
fprintf(file, "Major revision number: %d\n", devProp.major);
fprintf(file, "Minor revision number: %d\n", devProp.minor);
fprintf(file, "Name: %s\n", devProp.name);
fprintf(file, "Total global memory: %lu\n", (unsigned long) devProp.totalGlobalMem);
fprintf(file, "Total shared memory per block: %lu\n", (unsigned long) devProp.sharedMemPerBlock);
fprintf(file, "Total registers per block: %d\n", devProp.regsPerBlock);
fprintf(file, "Warp size: %d\n", devProp.warpSize);
fprintf(file, "Maximum memory pitch: %lu\n", (unsigned long) devProp.memPitch);
fprintf(file, "Maximum threads per block: %d\n", devProp.maxThreadsPerBlock);
for (int i = 0; i < 3; ++i)
fprintf(file, "Maximum dimension %d of block: %d\n", i, devProp.maxThreadsDim[i]);
for (int i = 0; i < 3; ++i)
fprintf(file, "Maximum dimension %d of grid: %d\n", i, devProp.maxGridSize[i]);
fprintf(file, "Clock rate: %d\n", devProp.clockRate);
fprintf(file, "Total constant memory: %lu\n", (unsigned long) devProp.totalConstMem);
fprintf(file, "Texture alignment: %lu\n", (unsigned long) devProp.textureAlignment);
fprintf(file, "Concurrent copy and execution: %s\n", (devProp.deviceOverlap ? "Yes" : "No"));
fprintf(file, "Number of multiprocessors: %d\n", devProp.multiProcessorCount);
fprintf(file, "Kernel execution timeout: %s\n", (devProp.kernelExecTimeoutEnabled ? "Yes" : "No"));
fflush(file);
return;
}
const cudaDeviceProp& getCurrentDeviceProperties() {
int device = 0;
auto err = cudaGetDevice(&device);
checkCuda(err, std::string("CUDA ERROR: cudaGetDeviceCount "));
return getDeviceProperties(device);
}
/**
* Returns the number of multiprocessors in the selected GPU.
* @return the number of multiprocessors in the selected GPU.
*/
int getGPUMultiprocessors() {
cudaDeviceProp prop = getDeviceProperties();
return prop.multiProcessorCount;
}
