参考博客:https://www.pianshen.com/article/943122806/
一、新建项目
打开VS2015 → 新建项目 → 其他 → “空项目”
二、调整配置管理器平台类型
右键项目属性 → 配置管理器 → 全改为“x64”
三、配置生成属性
右键项目 →生成依赖项 →生成自定义 →勾选“CUDA 9.2xxx”
四、配置基本库文件(CUDA9.2使用默认安装路径)
右键项目 → 属性 → 配置属性 → VC++目录 → 包含目录,添加以下目录:
- C:Program FilesNVIDIA GPU Computing ToolkitCUDAv9.2include
…… → 库目录,添加以下目录:
- C:Program FilesNVIDIA GPU Computing ToolkitCUDAv9.2libx64
五、配置CUDA静态链接库路径
右键项目 → 属性 → 配置属性 → 链接器 → 常规 → 附加库目录,添加以下目录:
- $(CUDA_PATH_V9_2)lib$(Platform)
六、选用CUDA静态链接库(直接复制即可)
右键项目 → 属性 → 配置属性 → 链接器 → 输入 → 附加依赖项,添加以下库:
- cublas.lib;cublas_device.lib;cuda.lib;cudadevrt.lib;cudart.lib;cudart_static.lib;cufft.lib;cufftw.lib;curand.lib;cusolver.lib;cusparse.lib;nppc.lib;nppial.lib;nppicc.lib;nppicom.lib;nppidei.lib;nppif.lib;nppig.lib;nppim.lib;nppist.lib;nppisu.lib;nppitc.lib;npps.lib;nvblas.lib;nvcuvid.lib;nvgraph.lib;nvml.lib;nvrtc.lib;OpenCL.lib;
以上为 “第三步” 中添加的库目录 “C:Program FilesNVIDIA GPU Computing ToolkitCUDAv9.0libx64” 中的
库!
七、配置源码文件风格
右键源文件 → 添加 → 新建项 → 选择 “CUDA C/C++ File”
右键 “xxx.cu" 源文件 → 属性 → 配置属性 → 常规 → 项类型 → 设置为“CUDA C/C++”
八 将测试程序拷贝到 .cu文件中
1 #include "cuda_runtime.h" 2 3 #include "device_launch_parameters.h" 4 5 #include <stdio.h> 6 #include <stdlib.h> 7 8 9 void main() { 10 11 int deviceCount; 12 13 cudaGetDeviceCount(&deviceCount); 14 15 16 17 int dev; 18 19 for (dev = 0; dev < deviceCount; dev++) 20 21 { 22 23 int driver_version(0), runtime_version(0); 24 25 cudaDeviceProp deviceProp; 26 27 cudaGetDeviceProperties(&deviceProp, dev); 28 29 if (dev == 0) 30 31 if (deviceProp.minor = 9999 && deviceProp.major == 9999) 32 33 printf(" "); 34 35 printf(" Device%d:"%s" ", dev, deviceProp.name); 36 37 cudaDriverGetVersion(&driver_version); 38 39 printf("CUDA驱动版本: %d.%d ", driver_version / 1000, (driver_version % 1000) / 10); 40 41 cudaRuntimeGetVersion(&runtime_version); 42 43 printf("CUDA运行时版本: %d.%d ", runtime_version / 1000, (runtime_version % 1000) / 10); 44 45 printf("设备计算能力: %d.%d ", deviceProp.major, deviceProp.minor); 46 47 printf("Total amount of Global Memory: %u bytes ", deviceProp.totalGlobalMem); 48 49 printf("Number of SMs: %d ", deviceProp.multiProcessorCount); 50 51 printf("Total amount of Constant Memory: %u bytes ", deviceProp.totalConstMem); 52 53 printf("Total amount of Shared Memory per block: %u bytes ", deviceProp.sharedMemPerBlock); 54 55 printf("Total number of registers available per block: %d ", deviceProp.regsPerBlock); 56 57 printf("Warp size: %d ", deviceProp.warpSize); 58 59 printf("Maximum number of threads per SM: %d ", deviceProp.maxThreadsPerMultiProcessor); 60 61 printf("Maximum number of threads per block: %d ", deviceProp.maxThreadsPerBlock); 62 63 printf("Maximum size of each dimension of a block: %d x %d x %d ", deviceProp.maxThreadsDim[0], 64 65 deviceProp.maxThreadsDim[1], 66 67 deviceProp.maxThreadsDim[2]); 68 69 printf("Maximum size of each dimension of a grid: %d x %d x %d ", deviceProp.maxGridSize[0], deviceProp.maxGridSize[1], deviceProp.maxGridSize[2]); 70 71 printf("Maximum memory pitch: %u bytes ", deviceProp.memPitch); 72 73 printf("Texture alignmemt: %u bytes ", deviceProp.texturePitchAlignment); 74 75 printf("Clock rate: %.2f GHz ", deviceProp.clockRate * 1e-6f); 76 77 printf("Memory Clock rate: %.0f MHz ", deviceProp.memoryClockRate * 1e-3f); 78 79 printf("Memory Bus Width: %d-bit ", deviceProp.memoryBusWidth); 80 81 } 82 system("pause"); 83 //return 0; 84 85 }
成功的结果: