Intel® VTune™ Amplifier XE and Intel® VTune™ Amplifier for Systems Help
OpenMP Potential Gain Creation shows elapsed time cost of parallel work arrangement by OpenMP runtime library. If the time is significant consider parallelizing outer looks rather than inner loops where possible.
CPU time spent on parallel work arrangement can be a result of too fine-grain parallelism. Try parallelizing outer loops, rather than inner loops, to reduce the work arrangement overhead.