This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
It would be immensely cool and greatly enhance the usefulness of this tool if one could save off to a file all the underlying data for all the graphs and profiling result views, so that one could later re-load this data and continue additional analysis.... or send off to someone else for help to analyze performance problems. Also it would allow for automated profiling runs to be done, a 'nightly profile', and the results to be saved to file so a human could come back, load the file and review the collected performance data.
The profiler in the Hyades project http://www.eclipse.org/hyades/ does the same thing, allows profiling information to be stored so it can be browsed with the UI at a later point in time as well. This has been a very useful capability.
We would like to change the way results are held in memory, which should allow a lot of things to be implemented: - more efficient "live data" display (currently implemented as periodic "Get results" - saving/reopening results data - ability to more efficiently browse/search the results - ability to more efficiently visualize the results (e.g. not collapse the trees, compute aggregated data only on demand when a tree is expanded, etc.)
Will be done in 1.0
Planned for M7. The internal implementation depends on issue 51274 implementation.
To clarify the dependency: issue 51274 should address the current problem of quick obtaining and economic storage of CPU profiling data for all profiled threads at once. This will require modification of the internal format for storing presentation-level CPU profiling data. This format is going to be a "flat" one, which will make it easy to save/retrieve from disk, which is a subject of this issue (51542).
Planned for M7, though likely for CPU profiling data only.
Done for CPU profiling data, as the one where capability to save/retrieve data seems most essential. Will probably need to open another issue for memory data snapshots, if it's demanded.
I'll keet this one open, as this is not specific for CPU results.
In progress for M7
Fixed in M7
Verification of old issues.
Closing old issues.