Modules are executable binary image files. The Module view displays all the modules within the selected thread(s) running during sampling data collection.
Use the information from the Module view to drill down to the hotspots. Modules that have been called frequently during sampling data collection have the highest number of events or the most processor time.
For event-based sampling, the Module view displays the event totals (for each event) for every module that executed.
For Java* applications, all the Java methods are combined and displayed as one horizontal bar in the Module view.
Double-click the bar to view the individual methods and drill down to the Hotspot view by double-clicking this bar.
Right-click the view, and select View As Table or View as Bar Chart from the pop-up menu.
You can click the column header to sort and display the data in an ascending or descending order.
Column |
Description |
---|---|
Module |
Modules within the selected threads for which data was
collected. |
<Event name> CPU<n> events |
Number of events that occurred on each processor for every module. Processor information on which the sampling data was
collected is displayed if the Hide/Show CPU Information |
<Event name> events |
Total event count for each module. The event count is based on the number of samples collected for the module x Sample After value |
Process |
Names of the processes that called each module. |
Process Path |
Paths of the processes that called each module during sampling data collection. |
Module Path |
Paths of the modules during sampling data collection. |
Vertical axis |
Displays the names of the software modules that were active on your system when the samples were collected, including applications activated by the VTune(TM) Performance Analyzer and active system resources, such as device drivers, concurrent tasks, and OS functions. |
Horizontal axis |
The horizontal axis is a log scale of the number of samples collected. Each bar next to a module represents an event. The length of the bar indicates the total events recorded for that module. The longer the bar, the larger the number of events recorded. If two or more events are selected for sampling, then each event is color-coded and displayed next to a module. Place your cursor over a bar to display a tool tip with the module name, event name, and the number of events collected for that module. To adjust the scale, right-click the view and select Log Scale. |
Modules with names such as Other86, Other16, Other32, or Other64 are modules whose real name is not determined from the sampling data during the Activity run. The "XX" in OtherXX refers to the addressing mode of the processor during sampling data collection. For example, 32-bit mode is translated to Other32, 16-bit mode to Other16 and so on. Drilling down on OtherXX modules takes you to the Hotspot view grouped by address. The Address column provides the Relative Virtual Address (RVA) where the sample is collected. Since the VTune analyzer is not able to associate this address with any known modules, hotspots and source views are not available. See Troubleshooting Sampling Performance Data for more information.
Modules that took the most processor time (see the Timer or the Clockticks event columns).
Modules that generated the highest event counts for specific microarchitecture level performance events.
Performance of your module in relation to others.
Identify the critical module or modules and drill down to the Hotspot view.
Double-click a single module.
OR
Select several modules.
Click Hotspot
in the sampling toolbar.
Click Show/Hide
CPU Information in the sampling toolbar to display the number of samples
collected for each item per processor on your system.
If you try to drill down to hotspots view and do not see function names for the module that you expect to see, one common reason is that symbol association for that module has not been performed.
Click the following buttons for more information about sampling: