Java Profiling Tool Using AspectJ

1
Java Profiling Tool Using AspectJ

• By
• Aaron Skinner

2
Outline

• Introduction
• Problem Addressed
• Solution
• Concerns
• Demonstration
• Conclusion

3
Why Profiling

• Catch problems in early stages of development
• Produce higher quality software
• Track Program Execution

4
Existing Profilers

• Almost all Java based
• Commercially based
• 1 Using AspectJ Glassbox Inspector

5
Considerations

• Use AspectJ to monitor execution times and memory
  usage
• Take advantage of new features in Java 1.5
• Latest stable AspectJ compiler is 1.2
• Pre-release compiler for version 1.5

6
Solution

• Use AspectJ to track execution times and memory
  execution
• Use Windows Application (GUI) to Create Aspect
  Re-compile
• Display Profiling Results

7
Solution (cont.)

• Generic Aspect for profiling
• Test Aspects with simple programs to ensure
  expected functionality
• User interface to form Aspect file based on input
  parameters

8
Available Memory Resources

• MemoryPoolMXBean
• OperatingSystemMXBean           
• GarbageCollectorMXBean           

9
Memory Usage

• Heap
• runtime data area from which uses memory for all
  objects, such as arrays and instances of classes
• Non-Heap
• runtime constants, method data, actual byte code
  needed for methods and constructors

10
Heap Usage

• Deals with object allocation
• Affected by poor programming code
• Expose with AspectJ to profile runtime memory
  usage by an application

11
Profiler Windows Application
12
Concerns

• Overhead of adding the aspects
• Taking advantage of new features not yet
  available
• Usage Threshold
• Memory Pools

13
Demonstration
14
Future Concerns

• Development of simple framework
• Tracking of Peak Usage

15
Conclusions

• See Detailed Runtime Data
• Find possible problems in code
• Time Saving / Quality Assurance