#pragma acc parallel loop //Memory movement handled automagically
for(int i=0;i<10000000;i++)
  a[i] = std::sqrt(std::log(a[i])+1.0); //Tada! I'm on the GPU