The following figure shows steps required for extending regular hardware event-based sampling analysis to collect call stacks, threading and power efficiency data. To display more information about a workflow step:
- Position (hover) the mouse pointer to display a brief explanation.
- Click to display the associated topic.
Note
If clicking the workflow step does not display the associated topic, use the text links below the figure.
Prerequisites: Make sure your kernel is configured to enable the event-based stack sampling collection.
|
1. |
Create a project and choose an analysis target using the Configure Project toolbar button. Event-based stack sampling data collection cannot be configured for the entire system. You have to specify an application to launch or attach to. |
|
|
2. |
In the Analysis Type window, choose an event-based sampling analysis type with hardware events that fit your task and intention best. |
|
|
3. |
In the configuration pane for the selected analysis, enable the Collect stacks option. If this option is enabled, you may also select the Estimate call counts checkbox to statistically approximate the number of calls to sampled functions. |
|
|
4. |
Click Start to launch the selected event-based sampling analysis type. |
|
|
5. |
Explore the collected data and analyze performance, parallelism and power efficiency metrics with regard to execution paths. |