According to Wilcox (2005), Local Optimization Technique is the act of inspecting the codes one line at a time to look for known inefficiencies. Intuition Technique is to focus on the part of the codes that the developer suspect is slow. However, these approaches have proved to be ineffective even with experienced developers. Hence, a profiling tool is required to perform the task.
The Java Interactive Profiler was selected due to the following features:
1. It is available for download for free
2. It is lightweight and performs, on average, 10 to 20 times faster than using hprof
3. It tracks and factors out the overhead associated with gathering performance data, thus providing more accurate performance timings.
The Java Interactive Profiler was used to perform a profile on Algorithm1. Optimization made based on the profile achieved a general improvement of between 16.4% and 18.2% in efficiency for the 5 problems tested.
The Java Interactive Profiler was also used to perform a profile on Algorithm2. Optimization made based on the profile resulted in an improvement of 17.7% in efficiency on the ftp06 problem.
These improved algorithms were subsequently used to perform the experiments in this project, reducing the amount of time required for each run of the algorithms.
