The CPU Call Tree and CPU Hot Spots views present cumulated data in tabular form where each line represents a segment of the call stacks of involved threads with regard to CPU consumption.
In the CPU table for each listed thread and method, certain data is associated arranged in columns. You can add and remove columns from the Choose Columns options of the pop-up menu or from mBProfiler's preferences. Go to Window >Preferences (or Eclipse >Preferences if you are using OS X) and then to mBProfiler > Columns.
Parameter |
Description |
|---|---|
name– |
The name of the thread or the method/stack frame. |
class |
The class which defines the method. If not shown, its value is added to the name column. If there are several Class objects with the same name created from different loaders in a profiled an OSGi runtime, e.g. if a bundle is re-installed a few times, mBProfiler will add sequence numbers next to the symbolic name of repeating nodes, e.g. my.bundle.Demo (2). |
package |
The package of the class where the method belongs. If not shown, its value is added in the name column. |
class loader |
The name of the class loader, which has loaded the corresponding method.
|
time (ms) |
Shows the following data in CPU Call Table:
|
own time (ms) |
Shows the following data:
This value does not imply the duration of its children. |
inv count |
The number of invocations of the methods (available only for profiling with the timing method). For CPU Hot Spots, inv count shows how many times the hot spot is called by the relevant stack frame or method. |
avg time (ms) |
The average time for invocation execution (available only for the timing method). |
Table Filtering
You can filter the information in the CPU table and concentrate only on certain set of data.
Table Sorting
The information in the CPU table can be sorted according to one or two columns.