Java Application CPU Profiling

How to do Java Application's CPU Profiling using Java VisualVM to figure out where most of the time is spent by a Java program.

CPU Profiling

CPU Profiler works by taking thread dumps of the system at some regular interval (if using Sampling Profiler) and reporting on where it observes time being spent. This will allow developers to focus on the parts in their applications where most of the execution time is spent and try to optimize such problem areas by taking appropriate corrective actions.

The sampling rate is configurable. Other option is Tracing Profiler but it may cause more performance overheads.

Using Java VisualVM For CPU Profiling

Path of Java VisualVM- <JAVA_HOME>/bin/jvisualvm.exe

Java VisualVM - Install Profiler Plugin

Open Java VisualVM and make sure to install Profiler Plugin from the Available Plugins tab. Once installed, The plugin should be listed in the Installed tab as shown in below screenshot-

Java VisualVM Settings

Now open the Profile Startup setting window from the Application -> Profile Startup menu.

Things to note-

• Use an available port number on your system for attaching the profiler to your app to be profiled
• Start Profiling From Classes section should your application classes or packages

Now copy the generated JVM parameters by clicking on Copy to Clipboard button.

Starting the Application By Enabling Profiling

Now go to Eclipse Run Configurations and paste the copied JVM arguments to the VM Arguments section as shown below; save and run the app.

If you get the error code 62, then do this-

CPU Profiling Fails With "Redefinition failed with error 62"
Resolution: Restart the application with the VM argument -Xverify:none to disable classfile verification. 

Profiling Results

After this app run terminates, you would be prompted by Java VisualVM to save the snapshot of this run. Save the snapshot and view it by selecting it. This should look as shown below-

Now you can see how much time was spent in each of the methods in the call tree.

Important Note

It is advised that you don't run profiling in production and other higher version as byte code instrumentation has some overhead. So, care must be taken to run profiling on your development workspace only.

Memory Profiling

This profiler can be used to do memory profiling as well simply by selecting Memory profiling while setting Application Profile options.