Related Topics:
Related Topics:| Clock-Based Profiling Data | Annotated Source and Disassembly Code |
| Callers-Callees Function Metrics | Hardware Counter Overflow Profiling Data |
In multithreaded or MPI programs, synchronization wait tracing keeps track of real-time delays in calls to thread-synchronization or MPI barrier routines if the delay exceeds a certain predetermined threshold. Each data point contains time stamps for the beginning and the end of the delay, and the IDs of the thread and LWP running at the end of the delay.
Synchronization wait tracing information is converted into the following metrics:
The synchronization delay information that appears in the Function List display shows, for each function, the count of events recorded and the total number of seconds over threshold spent waiting on calls to synchronization routines.
By default, if synchronization wait tracing data has been recorded, the Function List display shows the following:
The data is sorted on inclusive synchronization wait times if the only data recorded is synchronization wait tracing data. The defaults can be changed using a defaults file. See the er_print(1) man page for details.
You can choose to display metrics as a percentage of total program metric, select other metrics to display and specify a different sort order in the Function List Metrics Dialog Box. See Choosing the Metrics Displayed and the Sort Order for more information.
You can also select a subset of experiments, samples, threads and LWPs for which to examine metrics. See Selecting Experiments, Samples, Threads, and LWPs for Analysis.