Profiling Results - Code Fragment
When you analyze code fragment performance, the results in the
Profiling Results tab display the time it takes to execute an
piece of code within one method, or the whole method.
Code fragment performance is profiled when you choose the
Analyze Code Fragment Performance in the Select Profiling Task dialog box.
To open the Profiling Results tab, click the Live Results button
( 
) in the Profiler window.
The Profiling Results tab
When you analyze code fragment performance, the results appear as a list in the Profiling Results tab.
The length of this list is the smaller of the following two values:
- The number of times that
this code region has actually been executed
- The size of the code region
CPU results buffer.
How code fragment performance is measured
The CPU results buffer is a ring buffer kept in the target JVM memory;
by default its size is 1000 elements. The size can be changed in the Custom
Profiling mode. Note that when you shut down the target JVM, the results
contained in this buffer are lost, thus you should invoke choose Profile ->
Reset Collected Results while the target VM is still active (though the target application
could have completed execution by that time).
When measuring code fragment performance you measure the time it takes to execute an
arbitrary piece of code within one method, or the whole method. It is as if
you inserted two System.currentTimeMillis() calls into the code
of your target method (or more, if there are any return statements
within this piece of code) and then calculated the difference. Profiler, however,
uses a high-resolution timer and automates the whole process, allowing you to
insert/remove as many timers as you want while your application is running.
Only one timer can be active at any time, however.
See also
