CUDA プログラミング - Runtime API

#include <cuda_runtime.h>
#include <iostream>
 
int main() {
    int deviceCount;
    // 利用可能なデバイスの数を取得
    cudaGetDeviceCount(&deviceCount);
 
    for (int i = 0; i < deviceCount; ++i) {
        cudaDeviceProp prop;
        // 指定したデバイスのプロパティを取得し、cudaDeviceProp 構造体に格納
        cudaGetDeviceProperties(&prop, i);
 
        std::cout << "Device " << i << ": " << prop.name << "\n";
        std::cout << "  Total Global Memory: " << prop.totalGlobalMem / (1024 * 1024) << " MB\n";
        std::cout << "  Shared Memory per Block: " << prop.sharedMemPerBlock << " bytes\n";
        std::cout << "  Registers per Block: " << prop.regsPerBlock << "\n";
        std::cout << "  Warp Size: " << prop.warpSize << "\n";
        std::cout << "  Max Threads per Block: " << prop.maxThreadsPerBlock << "\n";
        std::cout << "  Max Threads Dimension: [" << prop.maxThreadsDim[0] << ", " << prop.maxThreadsDim[1] << ", " << prop.maxThreadsDim[2] << "]\n";
        std::cout << "  Max Grid Size: [" << prop.maxGridSize[0] << ", " << prop.maxGridSize[1] << ", " << prop.maxGridSize[2] << "]\n";
        std::cout << "  Clock Rate: " << prop.clockRate / 1000 << " MHz\n";
        std::cout << "  Multi Processor Count: " << prop.multiProcessorCount << "\n";
        std::cout << "  Compute Capability: " << prop.major << "." << prop.minor << "\n";
    }
 
    return 0;
}