错误：不允许从__global__函数调用__host__函数 [英] error: calling a __host__ function from a __global__ function is not allowed

问题描述

我写了cuda函数来对特征点进行密集采样，但是我遇到了错误。我的CUDA代码如下。我正在使用cuda 7.5工具包。

I have written cuda function for dense sampling of feature points but i am getting error. My cuda code is given below. I am using cuda 7.5 toolkit.

#include <stdlib.h>
#include <stdio.h>
#include <math.h>
#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
#include "opencv2/imgproc/imgproc.hpp"
#include <opencv2/gpu/gpu.hpp>
#include <opencv2/opencv.hpp>


using namespace cv::gpu;
using namespace cv;
using namespace std;

__global__ void densefun(std::vector<int>* d_counters,std::vector<Point2f>* d_points,int d_x_max,int d_y_max,int width, int min_distance)
{
  int i = blockDim.x * blockIdx.x + threadIdx.x;
  Point2f point = (*d_points)[i];
  int x = cvFloor(point.x);
  int y = cvFloor(point.y);
  //if(x >= d_x_max || y >= d_y_max)
      //continue;
  x /= min_distance;
  y /= min_distance;
  (*d_counters)[y*width+x]++;
}


void dense(std::vector<int>& counters,std::vector<Point2f>& points,int x_max,int y_max,int width)
{
  std::vector<int>* d_counters;
  std::vector<Point2f>* d_points;
  int min_distance=5; 
  cudaMalloc(&d_counters,counters.size());
  cudaMalloc(&d_points,points.size());
  cudaMemcpy(d_points, &points, points.size(), cudaMemcpyHostToDevice);
  densefun<<<1,points.size()>>>(d_counters,d_points,x_max,y_max,width,min_distance);
  cudaMemcpy(&counters, d_counters, counters.size(), cudaMemcpyDeviceToHost);
  cudaFree(d_counters);
  cudaFree(d_points);
}

输出：

/home/supriya/Desktop/5Dec/CalculateFV_merged_gpu_old/build/denseCuda.cu（28）：
错误：调用 host 函数（ cv ::
a global 函数（ densefun）中的Point_ :: Point_）不允许

/home/supriya/Desktop/5Dec/CalculateFV_merged_gpu_old/build/denseCuda.cu(28): error: calling a host function("cv::Point_ ::Point_") from a global function("densefun") is not allowed

/ home / supriya / Desktop / 5Dec / CalculateFV_merged_gpu_old / build / denseCuda.cu（28）：
错误：调用 host 函数（ std :: vector，
std :: allocator>> :: operator 全局
函数（ densefun）中的[]）不允许

/home/supriya/Desktop/5Dec/CalculateFV_merged_gpu_old/build/denseCuda.cu(28): error: calling a host function("std::vector , std::allocator > > ::operator []") from a global function("densefun") is not allowed

/ home / supriya / Desktop / 5Dec /CalculateFV_merged_gpu_old/build/denseCuda.cu(29）
（第7行）：错误：从
global <中调用 host 函数（ cvFloor） / strong>不允许使用功能（ densefun）

/home/supriya/Desktop/5Dec/CalculateFV_merged_gpu_old/build/denseCuda.cu(29) (col. 7): error: calling a host function("cvFloor") from a global function("densefun") is not allowed

/home/supriya/Desktop/5Dec/CalculateFV_merged_gpu_old/build/denseCuda.cu（30）
（第7条）：错误：呼叫主机 全局函数中的trong>函数（ cvFloor）

/home/supriya/Desktop/5Dec/CalculateFV_merged_gpu_old/build/denseCuda.cu(30) (col. 7): error: calling a host function("cvFloor") from a global function("densefun") is not allowed

/ home / supriya /Desktop/5Dec/CalculateFV_merged_gpu_old/build/denseCuda.cu(35）：
错误：从<调用 host 函数（ std :: vector> :: operator []）不允许strong> global
函数（ densefun）

/home/supriya/Desktop/5Dec/CalculateFV_merged_gpu_old/build/denseCuda.cu(35): error: calling a host function("std::vector > ::operator []") from a global function("densefun") is not allowed

在
/ tmp / tmpxft_00000c85_00000000-7_denseCuda.cpp1.ii。
的CMake错误testVideo_generation_denseCuda.cu.o.cmake：260（消息）：错误
生成文件

/ home / supriya / Desktop / 5Dec / CalculateFV_merged_gpu_old / build / CMakeFiles / testVideo .dir //./ testVideo_generation_denseCuda.cu.o

5 errors detected in the compilation of "/tmp/tmpxft_00000c85_00000000-7_denseCuda.cpp1.ii". CMake Error at testVideo_generated_denseCuda.cu.o.cmake:260 (message): Error generating file
/home/supriya/Desktop/5Dec/CalculateFV_merged_gpu_old/build/CMakeFiles/testVideo.dir//./testVideo_generated_denseCuda.cu.o

CMakeFiles / testVideo.dir / build.make：392：目标
的配方'CMakeFiles / testVideo。 dir /./ testVideo_generation_denseCuda.cu.o'失败
make [2]： *
[CMakeFiles / testVideo.dir /./ testVideo_generation_denseCuda.cu.o]错误
1 CMakeFiles / Makefile2：130：目标
'CMakeFiles / testVideo.dir / all'的配方制作失败[1]：*
[CMakeFiles / testVideo.dir / all]错误2制作文件：76 ：目标
'all'的配方失败：*** [全部]错误2

CMakeFiles/testVideo.dir/build.make:392: recipe for target 'CMakeFiles/testVideo.dir/./testVideo_generated_denseCuda.cu.o' failed make[2]: * [CMakeFiles/testVideo.dir/./testVideo_generated_denseCuda.cu.o] Error 1 CMakeFiles/Makefile2:130: recipe for target 'CMakeFiles/testVideo.dir/all' failed make[1]: * [CMakeFiles/testVideo.dir/all] Error 2 Makefile:76: recipe for target 'all' failed make: *** [all] Error 2

推荐答案

您不能在CUDA内核中使用C ++标准库，OpenCV或任何其他非CUDA特定的库。

You cannot use C++ standard library, OpenCV or any other non-CUDA specific library inside a CUDA kernel.

您需要使用原始指针指向设备上分配的数组，而不是 std :: vector code> Point2f ，您需要使用特定于CUDA的向量类型 float2 ，而不是 cvFloor 您需要使用 __ device__ floorf（）等等。

Instead of std::vector you need to use raw pointers to arrays allocated on the device, instead of Point2f you need to use CUDA specific vector type float2, instead of cvFloor you need to use __device__ ​ floorf() and so on.

这篇关于错误：不允许从__global__函数调用__host__函数的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！