Intel® VTune™ Amplifier XE and Intel® VTune™ Amplifier for Systems Help
OpenMP Potential Gain Scheduling shows elapsed time cost of work assignment for threads by OpenMP runtime library. If the time is significant, consider using coarse-grain work chunking.
Dynamic scheduling with small work chunks can cause increased overhead due to threads frequently returning to the scheduler for more work. Try increasing the chunk size to reduce this overhead.