Be sure that you have "-knob enable-stack-collection=true" added to your command line or "Call Stack" knob switched ON in GUI configuration.
BTW - the function that you are looking at is OpenMP runtime function that usually signals imbalance on barriers or thread waiting on serial part of execution. If you want more details on OpenMP parallel efficiency I would recommend to run either Threading or HPC Performance Characterization analysis. See https://software.intel.com/en-us/vtune-amplifier-cookbook-openmp-code-analysis-method for more details.
