CUDA和C++结合编程中错误异常erro C2059:语法错误:"<"

创建c++工程后，通过c++调用cuda的程序实现两个矩阵相加的过程。在代码实现过程中cuda的add<<<>>>()函数总是报错。错误信息为：error C2059:语法错误:"<"。

两个文件的代码如下：

cudaTest.cu文件下的内容：

#include <stdio.h>#include "cuda_runtime.h"#include "device_launch_parameters.h"#include "time.h"#include <stdlib.h>#include <cstdlib>#define datasize 50000inline void checkCudaErrors(cudaError err){if (cudaSuccess != err){fprintf(stderr, "CUDA Runtime API error:%s.\n", cudaGetErrorString(err));return;}}__global__ void add(int *a, int *b, int *c){int tid = blockIdx.x*blockDim.x + threadIdx.x;for (size_t k = 0; k < 500000; k++){c[tid] = a[tid] + b[tid];}}extern "C" int runtest(int *host_a, int *host_b, int *host_c){int *dev_a, *dev_b, *dev_c;checkCudaErrors(cudaMalloc((void**)&dev_a, sizeof(int)*datasize));checkCudaErrors(cudaMalloc((void**)&dev_b, sizeof(int)*datasize));checkCudaErrors(cudaMalloc((void**)&dev_c, sizeof(int)*datasize));checkCudaErrors(cudaMemcpy(dev_a, host_a, sizeof(int)*datasize, cudaMemcpyHostToDevice));//从主机复制到显卡内存checkCudaErrors(cudaMemcpy(dev_b, host_b, sizeof(int)*datasize, cudaMemcpyHostToDevice));//从主机复制到显卡内存add <<<datasize / 100, 100 >>> (dev_a, dev_b, dev_c);//调用显卡处理数据checkCudaErrors(cudaMemcpy(host_c,dev_c, sizeof(int)*datasize, cudaMemcpyDeviceToHost));//从显卡复制到主机内存cudaFree(dev_a);cudaFree(dev_b);cudaFree(dev_c);return 0;}extern "C" int main2(){int a[datasize], b[datasize], c[datasize];for (size_t i = 0; i < datasize; i++){a[i] = i;b[i] = i*i;}long now1 = clock();runtest(a, b, c);printf("GPU运行时间为：%dms\n", int(((double)(clock() - now1)) / CLOCKS_PER_SEC * 1000));long now2 = clock();for (size_t i = 0; i < datasize; i++){for (size_t k = 0; k < 50000; k++){c[i] = a[i] + b[i];}}printf("CPU运行时间为：%dms\n", int(((double)(clock() - now2)) / CLOCKS_PER_SEC * 1000));getchar();return 0;}

c++控制台应用程序文件为：ConsoleApplication.cpp文件

// ConsoleApplication1.cpp : 定义控制台应用程序的入口点。//#include "stdafx.h"#include "cudaTest.cu"extern "C" int main2();extern "C" int runtest(int *host_a, int *host_b, int *host_c);int _tmain(int argc, _TCHAR* argv[]){main2();return 0;}

通过上述调用过程，程序一直报错为：C2059错误，其中错误问题定位在add<<<>>>()函数上，如下所示：

解决方法，其实很简单，将ConsoleApplication.cpp文件中的头文件

#include "cudaTest.cu"

删掉即可。这样c++工程下的文件不包含cuda下的函数，只不过调用其中函数直接执行即可。ConsoleApplication.cpp文件的内容修改后为：

// ConsoleApplication1.cpp : 定义控制台应用程序的入口点。//#include "stdafx.h"extern "C" int main2();extern "C" int runtest(int *host_a, int *host_b, int *host_c);int _tmain(int argc, _TCHAR* argv[]){main2();return 0;}