c++ - OpenMP: how to handle race condition on global std vectors -

i'm interested in parallelize loop using openmp global std vectors involved.

in specific example, order in operation vec[1] = vec[1] + i executed doesn't matter results appears wrong due race conditions. proper way handle this?

the reduction clause seems not working containers.

#include <iostream> #include <vector>   std::vector<double> vec = {1,1};  void func(int i){     vec[1] = vec[1] + i; }  int main(int argc, char *argv[]) {     int k;     #pragma omp parallel     for(k=1; k<10; k++){         func(k);     }     std::cout << vec[0] << ", " << vec[1] << std::endl; } 

the problem line

vec[1] = vec[1] + i; 

isn't atomic. when compiled, looks more this

auto tmp = vec[1]; tmp = tmp + i; vec[1] = tmp 

and in case, have race condition.

if want way, can tell openmp vec[1] = vec[1] + i; should atomic, :

#include <iostream> #include <vector>  std::vector<double> vec = {1,1};  void func(int i){     #pragma omp atomic     vec[1] = vec[1] + 1; }  int main(int argc, char *argv[]) {     int k;     #pragma omp parallel     for(k=1; k<100; k++){         func(k);     }     std::cout << vec[0] << ", " << vec[1] << std::endl; }