In CUDA when many threads asking for the same data, it broadcast. Is the same thing happens also in openMP? Please help.