The Functions tab shows a list of functions and their metrics. Only the functions that have non-zero metrics are listed. The Functions tab can display inclusive metrics and exclusive metrics.
Time metrics are shown as seconds, presented to millisecond precision. Percentages are shown to a precision of 0.01%. If a metric value is precisely zero, its time and percentage is shown as "0." If the value is not exactly zero, but is smaller than the precision, its value is shown as "0.000" and its percentage as "0.00". Because of rounding, percentages may not sum to exactly 100%. Count metrics are shown as an integer count.
The function list is sorted by the data in one of the columns, allowing you to identify easily which functions have high metric values. The sort column header text is displayed in bold face and a triangle appears in the lower left corner of the column header. You can select the sort metric column by clicking its column header. Changing the sort metric in the Functions tab changes the sort metric in the Callers-Callees tab unless the sort metric in the Callers-Callees tab is an attributed metric.
If you don't want to see the functions from a particular load object, you can hide them by
choosing View Show/Hide Functions.
When you do so, the load object appears in the function list instead of the functions.
In Java mode, the functions list shows metrics against interpreted
JavaTM
methods, and any native methods called. In this mode, the artificial
function <no Java call stack recorded> indicates that the
JavaTM Virtual Machine (JVM) software did not report a Java call stack,
even though a Java program was running.
Expert-Java mode additionally lists
methods that were dynamically compiled by the Java HotSpotTM
virtual machine. The Machine mode
shows multiple HotSpot compilations as completely independent functions, although the
functions will have the same name. In this mode, all functions from the JVM software
are shown as such.
To change the Java mode, choose View
Set Data Presentation...
Formats
and select the radio button for the desired Java mode.
Note: the terms "Java Virtual Machine" or "JVM" mean a virtual machine for the Java platform.
The metrics initially shown are based on the data collected and on the default settings read from various .er.rc files. The following table summarizes the default settings.
Data Type | Default Metrics |
---|---|
Clock-based profiling | inclusive and exclusive User CPU time |
Hardware counter overflow profiling | inclusive and exclusive times (for counters that count in cycles) or event counts (for other counters) |
Synchronization delay tracing | inclusive synchronization wait count and inclusive synchronization delay time |
Heap tracing | inclusive leaks and inclusive bytes leaked |
MPI tracing | inclusive MPI Time, inclusive MPI Bytes Sent, inclusive MPI Sends, inclusive MPI Bytes Received, inclusive MPI Receives, and inclusive MPI Other |
![]() |
You can repeat this procedure for different experiment selections. |
When you click the Compose Filter Clause button in the toolbar, the Filter clause text box in the Advanced tab of the Filter Data dialog box is loaded with a filter clause that reflects the selection(s) in the Functions tab.
![]() |
When you select multiple functions, the Compose Filter Clause button is added to the toolbar. When you click this button, the Filter dialog box opens with Advanced tab selected. |
See also | |
---|---|
Setting Data Presentation Preferences Filtering the Data Showing and Hiding Functions Exclusive, Inclusive, and Attributed Metrics Performance Data and its Interpretation Program Structures |