0

我在理解如何使用内核循环 3 维数组时遇到了一些问题。

这是我到目前为止的代码:

#include <iostream>
#include <ctime>

#include <cuda.h>
#include <cuda_runtime.h>
#include <device_launch_parameters.h>

using namespace std;


int main()
{
// Array properties
const int width = 1;
const int height = 1;
const int depth = 1;

// Declaration of arrays
float h_A[width][height][depth];
float h_B[width][height][depth];
float h_C[width][height][depth] = {{{0}}};

// Fill up arrays
srand(time(0));
for(int i = 0; i < width; i++){
    for(int j = 0; j < height; j++){
        for(int z = 0; z < depth; z++){
            h_A[i][j][z] = rand()%1000;
            h_B[i][j][z] = rand()%1000;
        }
    }
}

// Declaration of device pointers
cudaPitchedPtr d_A, d_B, d_C;

// Allocating memory in GPU
cudaExtent extent = make_cudaExtent(width*sizeof(float),height,depth);
cudaMalloc3D(&d_A, extent);
cudaMalloc3D(&d_B, extent);
cudaMalloc3D(&d_C, extent);

// Copying memory from host to device
cudaMemcpy3DParms p;
p.srcPtr = make_cudaPitchedPtr(&h_A, sizeof(float)*width, height, depth);
p.extent = extent;
p.kind = cudaMemcpyHostToDevice;

p.dstPtr = d_A;
cudaMemcpy3D(&p);
p.dstPtr = d_B;
cudaMemcpy3D(&p);
p.dstPtr = d_C;
cudaMemcpy3D(&p);

system("pause");
return 0;
}

如何制作一个循环遍历数组中每个元素并将它们添加在一起的内核?

4

1 回答 1

2

在 CUDA 4.0 编程指南的第 21 页上有一个示例,用于循环 2D 浮点数组:

// Host code
int width = 64, height = 64;
float* devPtr;
size_t pitch;
cudaMallocPitch(&devPtr, &pitch,
width * sizeof(float), height);
MyKernel<<<100, 512>>>(devPtr, pitch, width, height);


// Device code
__global__ void MyKernel(float* devPtr, size_t pitch, int width, int height)
{
   for (int r = 0; r < height; ++r) 
    {
       float* row = (float*)((char*)devPtr + r * pitch);
          for (int c = 0; c < width; ++c) 
              {
              float element = row[c];
              }
     }
}

重写它以总结元素应该很容易。此外,您可以参考线程。当关注效率时,您还可以查看 CUDA 中的并行缩减方法。例如,在实施 Monte Carlo 模拟时使用此功能(请参阅 Multi Monte Carlo 示例)。

于 2013-09-15T18:39:23.320 回答