profiler 5.0 help
Snapshots
A snapshot is a fully static snapshot of profiling results at the moment the snapshot is taken. Snapshots complement Live Results in that they contain a more detailed record of profiling data at a certain point in time.
Snapshots are displayed in the Source Editor when you click Take Snapshot in the Profiler window or when you click Take Snapshot of Profiling Results in the toolbar of the Profiling Results tab.
Saving Snapshots
After you have taken a snapshot, the controls in the snapshot toolbar enable you to save the snapshot files.
 |
Save Snapshot to Project
Click this button to save the snapshot to your project. The snapshot is saved to the profiler directory in nbproject > private directory in your project. Once saved, the snapshot files are visible in the Snapshots section of the Profiler window. |
 |
Save Snapshot to Custom File
Click this button to open a save dialog box to save the snapshot file in a location other than to your project. |
Additionally, the toolbar for CPU snapshots has a Method/Class/Package combo box which enables you to control the coarseness of the presentation of the results.
Saved Snapshots
The Saved Snapshots section in the Profiler Window enables you to manage your saved snapshots.
Snapshots are organized according to
project. When you select a project in the combo box, the snapshots saved
to that project are listed below the combo box.
The buttons to the right of the listed snapshots enable you to open and
delete your saved snapshots. The Save As button enables you to save the selected snapshot
to another location outside the project. The Load button enables you to locate
snapshots saved outside your project.
Snapshot are displayed according to the time the snapshot was taken. The data stored in a snapshot varies according to the type of profiling task. The following icons identify the type of snapshot.
 |
Memory snapshot |
 |
CPU snapshot |
 |
Code fragment snapshot |
Advantages of Taking Snapshots
Previously, when Get Results was used, only partial results were obtained from the profiled VM, and subsequent actions, such as selecting thread for CPU results and displaying allocation stack traces lead to further communication with the profiled application. This approach had two disadvantages:
- the profiled application needed to run during inspection of results
- the allocation stack traces and thread CPU results were not representing the time when Get Results was pressed, but the time when the thread was selected or allocation stack traces action invoked.
See also
