CUDA 6.5 && VS2013 && Win7：创建CUDA项目

$(CUDA_INC_PATH)

$(CUDA_PATH)/lib/$(PlatformName)

cudart.lib

1 #include "cuda.h"
2 #include "cuda_runtime.h"
3 #include "device_launch_parameters.h"

 1 #include "cuda.h" 
 2 #include "cuda_runtime.h"
 3 #include "device_launch_parameters.h"
 4 #include <stdio.h>
 5 
 6 __global__ void SaXPY(float a, float* X_d, float* Y_d, int n)
 7 {
 8     if (threadIdx.x < n)
 9         Y_d[threadIdx.x] = a * X_d[threadIdx.x] + Y_d[threadIdx.x];
10 }
11 
12 int main()
13 {
14     int n = 64;
15     float a = 2;
16     float *X_h, *X_d, *Y_h, *Y_d;
17     X_h = (float*)malloc(n * sizeof(float));
18     Y_h = (float*)malloc(n * sizeof(float));
19     for (int i = 0; i < n; i++)
20     {
21         X_h[i] = (float)i;
22         Y_h[i] = 1.0;
23     }
24     cudaMalloc(&X_d, n * sizeof(float));
25     cudaMalloc(&Y_d, n * sizeof(float));
26     cudaMemcpy(X_d, X_h, n * sizeof(float), cudaMemcpyHostToDevice);
27     cudaMemcpy(Y_d, Y_h, n * sizeof(float), cudaMemcpyHostToDevice);
28     SaXPY <<<1, 64 >>>(a, X_d, Y_d, n);
29     cudaMemcpy(Y_h, Y_d, n * sizeof(float), cudaMemcpyDeviceToHost);
30     for (int i = 0; i < n; i++)
31         printf("%2.1f X[%d] + Y[%d] = %f\n", a, i, i, Y_h[i]);
32     cudaFree(X_d);
33     cudaFree(Y_d);
34     free(X_h);
35     free(Y_h);
36     system("Pause");
37     return 0;
38 }