这两天因工作需要,写了个算法,计算复杂度较高。期望利用多核机器的优势,并行计算,加快运行速度。简单看了看资料,尝试两个小程序。
程序一:函数能并行运行就OK。
DWORD WINAPI Fun(LPVOID lpParamter) { cout<<"Fun display"endl; i++; } int main (void) { cout << "Now starting to verify multi-thread on windows" << endl; HANDLE hThread = CreateThread(NULL, 0, Fun, NULL, 0, NULL); CloseHandle(hThread); while(1) { cout<<"main display!"<<endl; Sleep(2000); } cout << "Press anykey to bye-bye" << endl; getchar(); return 0; }
程序二:传个参数进到并发函数中,如下:
DWORD WINAPI Fun(LPVOID lpParamter) { int iCount = *((int*)lpParamter); int i = 0; while(i < iCount) { cout<<"Fun display in the "<< i << " times" <<endl; i++; // Sleep(1000); } return iCount; } int main (void) { cout << "Now starting to verify multi-thread on windows" << endl; int i = 1000; HANDLE hThread = CreateThread(NULL, 0, Fun, &i, 0, NULL); CloseHandle(hThread); while(1) { cout<<"main display!"<<endl; Sleep(2000); } getchar(); return 0; }
程序三:在写上两个程序的时候,听说OpenMP神器,尝试一下
int main (void) { cout << "Now starting to verify multi-thread on windows" << endl; int iArray [100]; for (int i=0; i<100; i++) iArray[i] = i; #pragma omp parallel for num_threads(2) for (int i=0; i<100; i++) cout << iArray[i] << endl; getchar(); return 0; }
运行的时候,可以看到:1. 双核都在计算;2. 数组实际上是乱序输出(多个线程并发的结果)。
最后我工作中采用的是OpenMP。