Intel® VTune™ Amplifier
When the analysis run is complete, the Intel® VTune™ Amplifier generates a result is automatically opened in the default viewpoint. The location of the result files is specified in the Analysis Target tab.
A viewpoint typically contains the following elements:
|
Format |
Description |
|---|---|
|
Result Tab |
This is a container of all other viewpoint elements. This tab has the same name as the VTune Amplifier result file. The result tab name uses the r@@@{at} format, where @@@ is an incremented result number starting with 000 and at is the analysis type. For example: r003hs is the forth result run in this project and provides data for the Hotspot (hs) analysis. The Hotspots is the analysis type name. Hotspots by CPU Usage is the name of the viewpoint selected via the (change) link. Use this link to switch to other viewpoints available for this analysis type. |
|
Windows |
Each result tab includes a number of windows presenting colleted data from different perspectives. Each window has a corresponding tab. To ease your navigation, some windows are synchronized: when you select an element in a window, the same element is automatically selected in other windows of the same viewpoint. The list of windows depends on the selected viewpoint. Each window has a corresponding help topic available via F1 button. |
|
Panes |
Each window typically includes two or three panes, such as Call Stack pane, Timeline pane, Source pane, Assembly pane, and others. |
For a brief overview on a particular viewpoint, click the question mark icon at the viewpoint name.
All the viewpoints, regardless of the analysis type, make data display more convenient and manageable by:
Synchronizing Bottom-up, Top-down Tree windows, Timeline, and Call Stack panes to easily identify a program unit in review and the data related to this particular unit
Using the common criterion, defined by the Data of Interest, to calculate and present basic performance values (for example, CPU time for hotspot navigation in the source and calculation of the stack contribution)
Providing an option to choose a suitable data format (for example, colored bars or percent)
Providing a double-click mechanism to open the source code for a selected program unit
Providing a set of analysis specific filters to focus only on performance-critical program units
Introducing a selection mechanism to view total values for multiple program units
Providing an option to group data according to your preferences