Program Profiling & Optimization

full score

This project involves profiling and optimizing a target program using the gprof tool in a Linux environment. The process includes multiple stages of optimization based on profiling results.

There are two files in this project:

• bigram_og.c: The original target program to be optimized.
• bigram_opt.c: The optimized version of the target program. side note: Wrapper functions were used to check the time each function takes to execute.

• Perform multi-step program optimization and analysis through profiling.
• Use a self-developed target program, ideally the "Bigram Analyzer" studied in SP05-3 or any program with at least 15 functions, over 300 lines of code, and an initial runtime of over 3 minutes.

1. Implement and Test Target Program: Develop and verify the functionality of your target program.
2. Initial Profiling: Use gprof to perform the first profiling and analyze the results.
3. Select Functions for Optimization: Identify functions to optimize based on profiling analysis.
4. Perform Code Optimization: Optimize the selected functions.
5. Repeat Profiling and Optimization: Repeat steps 2-4 until sufficient optimization is achieved.
6. Conclusion: Finalize the optimization process and document the results.

• Focus on the systematic analysis and optimization process rather than the extent of performance improvement.
• Comprehensive documentation and analysis in the report are crucial.
• Any form of cheating will result in a score of zero and an "F" grade.